public-mirrors/linux
1
0
Fork 0
mirror of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git synced 2025-04-13 09:59:31 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
linux/fs/netfs/read_retry.c

314 lines
9.3 KiB
C
Raw Normal View History

netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
// SPDX-License-Identifier: GPL-2.0-only
/* Network filesystem read subrequest retrying.
*
* Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#include <linux/fs.h>
#include <linux/slab.h>
#include "internal.h"
static void netfs_reissue_read(struct netfs_io_request *rreq,
struct netfs_io_subrequest *subreq)
{
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
__clear_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
__set_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
netfs: Add retry stat counters Add stat counters to count the number of request and subrequest retries and display them in /proc/fs/netfs/stats. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-3-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:24:00 +00:00
netfs_stat(&netfs_n_rh_retry_read_subreq);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
subreq->rreq->netfs_ops->issue_read(subreq);
}
/*
* Go through the list of failed/short reads, retrying all retryable ones. We
* need to switch failed cache reads to network downloads.
*/
static void netfs_retry_read_subrequests(struct netfs_io_request *rreq)
{
struct netfs_io_subrequest *subreq;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
struct netfs_io_stream *stream = &rreq->io_streams[0];
struct list_head *next;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
_enter("R=%x", rreq->debug_id);
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
if (list_empty(&stream->subrequests))
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
return;
if (rreq->netfs_ops->retry_request)
rreq->netfs_ops->retry_request(rreq, NULL);
/* If there's no renegotiation to do, just resend each retryable subreq
* up to the first permanently failed one.
*/
if (!rreq->netfs_ops->prepare_read &&
netfs: Fix is-caching check in read-retry netfs: Fix is-caching check in read-retry The read-retry code checks the NETFS_RREQ_COPY_TO_CACHE flag to determine if there might be failed reads from the cache that need turning into reads from the server, with the intention of skipping the complicated part if it can. The code that set the flag, however, got lost during the read-side rewrite. Fix the check to see if the cache_resources are valid instead. The flag can then be removed. Fixes: ee4cdf7ba857 ("netfs: Speed up buffered reading") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/3752048.1734381285@warthog.procyon.org.uk cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:34:45 +00:00
!rreq->cache_resources.ops) {
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
break;
if (__test_and_clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags)) {
netfs: Work around recursion by abandoning retry if nothing read syzkaller reported recursion with a loop of three calls (netfs_rreq_assess, netfs_retry_reads and netfs_rreq_terminated) hitting the limit of the stack during an unbuffered or direct I/O read. There are a number of issues: (1) There is no limit on the number of retries. (2) A subrequest is supposed to be abandoned if it does not transfer anything (NETFS_SREQ_NO_PROGRESS), but that isn't checked under all circumstances. (3) The actual root cause, which is this: if (atomic_dec_and_test(&rreq->nr_outstanding)) netfs_rreq_terminated(rreq, ...); When we do a retry, we bump the rreq->nr_outstanding counter to prevent the final cleanup phase running before we've finished dispatching the retries. The problem is if we hit 0, we have to do the cleanup phase - but we're in the cleanup phase and end up repeating the retry cycle, hence the recursion. Work around the problem by limiting the number of retries. This is based on Lizhi Xu's patch[1], and makes the following changes: (1) Replace NETFS_SREQ_NO_PROGRESS with NETFS_SREQ_MADE_PROGRESS and make the filesystem set it if it managed to read or write at least one byte of data. Clear this bit before issuing a subrequest. (2) Add a ->retry_count member to the subrequest and increment it any time we do a retry. (3) Remove the NETFS_SREQ_RETRYING flag as it is superfluous with ->retry_count. If the latter is non-zero, we're doing a retry. (4) Abandon a subrequest if retry_count is non-zero and we made no progress. (5) Use ->retry_count in both the write-side and the read-size. [?] Question: Should I set a hard limit on retry_count in both read and write? Say it hits 50, we always abandon it. The problem is that these changes only mitigate the issue. As long as it made at least one byte of progress, the recursion is still an issue. This patch mitigates the problem, but does not fix the underlying cause. I have patches that will do that, but it's an intrusive fix that's currently pending for the next merge window. The oops generated by KASAN looks something like: BUG: TASK stack guard page was hit at ffffc9000482ff48 (stack is ffffc90004830000..ffffc90004838000) Oops: stack guard page: 0000 [#1] PREEMPT SMP KASAN NOPTI ... RIP: 0010:mark_lock+0x25/0xc60 kernel/locking/lockdep.c:4686 ... mark_usage kernel/locking/lockdep.c:4646 [inline] __lock_acquire+0x906/0x3ce0 kernel/locking/lockdep.c:5156 lock_acquire.part.0+0x11b/0x380 kernel/locking/lockdep.c:5825 local_lock_acquire include/linux/local_lock_internal.h:29 [inline] ___slab_alloc+0x123/0x1880 mm/slub.c:3695 __slab_alloc.constprop.0+0x56/0xb0 mm/slub.c:3908 __slab_alloc_node mm/slub.c:3961 [inline] slab_alloc_node mm/slub.c:4122 [inline] kmem_cache_alloc_noprof+0x2a7/0x2f0 mm/slub.c:4141 radix_tree_node_alloc.constprop.0+0x1e8/0x350 lib/radix-tree.c:253 idr_get_free+0x528/0xa40 lib/radix-tree.c:1506 idr_alloc_u32+0x191/0x2f0 lib/idr.c:46 idr_alloc+0xc1/0x130 lib/idr.c:87 p9_tag_alloc+0x394/0x870 net/9p/client.c:321 p9_client_prepare_req+0x19f/0x4d0 net/9p/client.c:644 p9_client_zc_rpc.constprop.0+0x105/0x880 net/9p/client.c:793 p9_client_read_once+0x443/0x820 net/9p/client.c:1570 p9_client_read+0x13f/0x1b0 net/9p/client.c:1534 v9fs_issue_read+0x115/0x310 fs/9p/vfs_addr.c:74 netfs_retry_read_subrequests fs/netfs/read_retry.c:60 [inline] netfs_retry_reads+0x153a/0x1d00 fs/netfs/read_retry.c:232 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 ... netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_dispatch_unbuffered_reads fs/netfs/direct_read.c:103 [inline] netfs_unbuffered_read fs/netfs/direct_read.c:127 [inline] netfs_unbuffered_read_iter_locked+0x12f6/0x19b0 fs/netfs/direct_read.c:221 netfs_unbuffered_read_iter+0xc5/0x100 fs/netfs/direct_read.c:256 v9fs_file_read_iter+0xbf/0x100 fs/9p/vfs_file.c:361 do_iter_readv_writev+0x614/0x7f0 fs/read_write.c:832 vfs_readv+0x4cf/0x890 fs/read_write.c:1025 do_preadv fs/read_write.c:1142 [inline] __do_sys_preadv fs/read_write.c:1192 [inline] __se_sys_preadv fs/read_write.c:1187 [inline] __x64_sys_preadv+0x22d/0x310 fs/read_write.c:1187 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83 Fixes: ee4cdf7ba857 ("netfs: Speed up buffered reading") Closes: https://syzkaller.appspot.com/bug?extid=1fc6f64c40a9d143cfb6 Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241108034020.3695718-1-lizhi.xu@windriver.com/ [1] Link: https://lore.kernel.org/r/20241213135013.2964079-9-dhowells@redhat.com Tested-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com Suggested-by: Lizhi Xu <lizhi.xu@windriver.com> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Reported-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-13 13:50:08 +00:00
__clear_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags);
subreq->retry_count++;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs_reset_iter(subreq);
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs_reissue_read(rreq, subreq);
}
}
return;
}
/* Okay, we need to renegotiate all the download requests and flip any
* failed cache reads over to being download requests and negotiate
* those also. All fully successful subreqs have been removed from the
* list and any spare data from those has been donated.
*
* What we do is decant the list and rebuild it one subreq at a time so
* that we don't end up with donations jumping over a gap we're busy
* populating with smaller subrequests. In the event that the subreq
* we just launched finishes before we insert the next subreq, it'll
* fill in rreq->prev_donated instead.
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
*
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
* Note: Alternatively, we could split the tail subrequest right before
* we reissue it and fix up the donations under lock.
*/
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
next = stream->subrequests.next;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
do {
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
struct netfs_io_subrequest *from, *to, *tmp;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
struct iov_iter source;
unsigned long long start, len;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
size_t part;
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
bool boundary = false, subreq_superfluous = false;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
/* Go through the subreqs and find the next span of contiguous
* buffer that we then rejig (cifs, for example, needs the
* rsize renegotiating) and reissue.
*/
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
from = list_entry(next, struct netfs_io_subrequest, rreq_link);
to = from;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
start = from->start + from->transferred;
len = from->len - from->transferred;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
_debug("from R=%08x[%x] s=%llx ctl=%zx/%zx",
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
rreq->debug_id, from->debug_index,
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
from->start, from->transferred, from->len);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
if (test_bit(NETFS_SREQ_FAILED, &from->flags) ||
!test_bit(NETFS_SREQ_NEED_RETRY, &from->flags))
goto abandon;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
list_for_each_continue(next, &stream->subrequests) {
subreq = list_entry(next, struct netfs_io_subrequest, rreq_link);
if (subreq->start + subreq->transferred != start + len ||
test_bit(NETFS_SREQ_BOUNDARY, &subreq->flags) ||
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
!test_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags))
break;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
to = subreq;
len += to->len;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
}
_debug(" - range: %llx-%llx %llx", start, start + len - 1, len);
/* Determine the set of buffers we're going to use. Each
* subreq gets a subset of a single overall contiguous buffer.
*/
netfs_reset_iter(from);
source = from->io_iter;
source.count = len;
/* Work through the sublist. */
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
subreq = from;
list_for_each_entry_from(subreq, &stream->subrequests, rreq_link) {
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
if (!len) {
subreq_superfluous = true;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
break;
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
}
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
subreq->source = NETFS_DOWNLOAD_FROM_SERVER;
subreq->start = start - subreq->transferred;
subreq->len = len + subreq->transferred;
__clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
netfs: Work around recursion by abandoning retry if nothing read syzkaller reported recursion with a loop of three calls (netfs_rreq_assess, netfs_retry_reads and netfs_rreq_terminated) hitting the limit of the stack during an unbuffered or direct I/O read. There are a number of issues: (1) There is no limit on the number of retries. (2) A subrequest is supposed to be abandoned if it does not transfer anything (NETFS_SREQ_NO_PROGRESS), but that isn't checked under all circumstances. (3) The actual root cause, which is this: if (atomic_dec_and_test(&rreq->nr_outstanding)) netfs_rreq_terminated(rreq, ...); When we do a retry, we bump the rreq->nr_outstanding counter to prevent the final cleanup phase running before we've finished dispatching the retries. The problem is if we hit 0, we have to do the cleanup phase - but we're in the cleanup phase and end up repeating the retry cycle, hence the recursion. Work around the problem by limiting the number of retries. This is based on Lizhi Xu's patch[1], and makes the following changes: (1) Replace NETFS_SREQ_NO_PROGRESS with NETFS_SREQ_MADE_PROGRESS and make the filesystem set it if it managed to read or write at least one byte of data. Clear this bit before issuing a subrequest. (2) Add a ->retry_count member to the subrequest and increment it any time we do a retry. (3) Remove the NETFS_SREQ_RETRYING flag as it is superfluous with ->retry_count. If the latter is non-zero, we're doing a retry. (4) Abandon a subrequest if retry_count is non-zero and we made no progress. (5) Use ->retry_count in both the write-side and the read-size. [?] Question: Should I set a hard limit on retry_count in both read and write? Say it hits 50, we always abandon it. The problem is that these changes only mitigate the issue. As long as it made at least one byte of progress, the recursion is still an issue. This patch mitigates the problem, but does not fix the underlying cause. I have patches that will do that, but it's an intrusive fix that's currently pending for the next merge window. The oops generated by KASAN looks something like: BUG: TASK stack guard page was hit at ffffc9000482ff48 (stack is ffffc90004830000..ffffc90004838000) Oops: stack guard page: 0000 [#1] PREEMPT SMP KASAN NOPTI ... RIP: 0010:mark_lock+0x25/0xc60 kernel/locking/lockdep.c:4686 ... mark_usage kernel/locking/lockdep.c:4646 [inline] __lock_acquire+0x906/0x3ce0 kernel/locking/lockdep.c:5156 lock_acquire.part.0+0x11b/0x380 kernel/locking/lockdep.c:5825 local_lock_acquire include/linux/local_lock_internal.h:29 [inline] ___slab_alloc+0x123/0x1880 mm/slub.c:3695 __slab_alloc.constprop.0+0x56/0xb0 mm/slub.c:3908 __slab_alloc_node mm/slub.c:3961 [inline] slab_alloc_node mm/slub.c:4122 [inline] kmem_cache_alloc_noprof+0x2a7/0x2f0 mm/slub.c:4141 radix_tree_node_alloc.constprop.0+0x1e8/0x350 lib/radix-tree.c:253 idr_get_free+0x528/0xa40 lib/radix-tree.c:1506 idr_alloc_u32+0x191/0x2f0 lib/idr.c:46 idr_alloc+0xc1/0x130 lib/idr.c:87 p9_tag_alloc+0x394/0x870 net/9p/client.c:321 p9_client_prepare_req+0x19f/0x4d0 net/9p/client.c:644 p9_client_zc_rpc.constprop.0+0x105/0x880 net/9p/client.c:793 p9_client_read_once+0x443/0x820 net/9p/client.c:1570 p9_client_read+0x13f/0x1b0 net/9p/client.c:1534 v9fs_issue_read+0x115/0x310 fs/9p/vfs_addr.c:74 netfs_retry_read_subrequests fs/netfs/read_retry.c:60 [inline] netfs_retry_reads+0x153a/0x1d00 fs/netfs/read_retry.c:232 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 ... netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235 netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371 netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407 netfs_dispatch_unbuffered_reads fs/netfs/direct_read.c:103 [inline] netfs_unbuffered_read fs/netfs/direct_read.c:127 [inline] netfs_unbuffered_read_iter_locked+0x12f6/0x19b0 fs/netfs/direct_read.c:221 netfs_unbuffered_read_iter+0xc5/0x100 fs/netfs/direct_read.c:256 v9fs_file_read_iter+0xbf/0x100 fs/9p/vfs_file.c:361 do_iter_readv_writev+0x614/0x7f0 fs/read_write.c:832 vfs_readv+0x4cf/0x890 fs/read_write.c:1025 do_preadv fs/read_write.c:1142 [inline] __do_sys_preadv fs/read_write.c:1192 [inline] __se_sys_preadv fs/read_write.c:1187 [inline] __x64_sys_preadv+0x22d/0x310 fs/read_write.c:1187 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83 Fixes: ee4cdf7ba857 ("netfs: Speed up buffered reading") Closes: https://syzkaller.appspot.com/bug?extid=1fc6f64c40a9d143cfb6 Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241108034020.3695718-1-lizhi.xu@windriver.com/ [1] Link: https://lore.kernel.org/r/20241213135013.2964079-9-dhowells@redhat.com Tested-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com Suggested-by: Lizhi Xu <lizhi.xu@windriver.com> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Reported-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-13 13:50:08 +00:00
__clear_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags);
subreq->retry_count++;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
/* Renegotiate max_len (rsize) */
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
stream->sreq_max_len = subreq->len;
netfs: Fix read-retry for fs with no ->prepare_read() Fix netfslib's read-retry to only call ->prepare_read() in the backing filesystem such a function is provided. We can get to this point if a there's an active cache as failed reads from the cache need negotiating with the server instead. Fixes: ee4cdf7ba857 ("netfs: Speed up buffered reading") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/529329.1736261010@warthog.procyon.org.uk cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-01-07 14:43:30 +00:00
if (rreq->netfs_ops->prepare_read &&
rreq->netfs_ops->prepare_read(subreq) < 0) {
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
trace_netfs_sreq(subreq, netfs_sreq_trace_reprep_failed);
__set_bit(NETFS_SREQ_FAILED, &subreq->flags);
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
goto abandon;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
}
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
part = umin(len, stream->sreq_max_len);
if (unlikely(stream->sreq_max_segs))
part = netfs_limit_iter(&source, 0, part, stream->sreq_max_segs);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
subreq->len = subreq->transferred + part;
subreq->io_iter = source;
iov_iter_truncate(&subreq->io_iter, part);
iov_iter_advance(&source, part);
len -= part;
start += part;
if (!len) {
if (boundary)
__set_bit(NETFS_SREQ_BOUNDARY, &subreq->flags);
} else {
__clear_bit(NETFS_SREQ_BOUNDARY, &subreq->flags);
}
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs_reissue_read(rreq, subreq);
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
if (subreq == to) {
subreq_superfluous = false;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
break;
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
}
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
}
/* If we managed to use fewer subreqs, we can discard the
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
* excess; if we used the same number, then we're done.
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
*/
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
if (!len) {
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
if (!subreq_superfluous)
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
continue;
list_for_each_entry_safe_from(subreq, tmp,
&stream->subrequests, rreq_link) {
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
trace_netfs_sreq(subreq, netfs_sreq_trace_superfluous);
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
list_del(&subreq->rreq_link);
netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_done);
if (subreq == to)
break;
}
continue;
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
}
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
/* We ran out of subrequests, so we need to allocate some more
* and insert them after.
*/
do {
subreq = netfs_alloc_subrequest(rreq);
if (!subreq) {
subreq = to;
goto abandon_after;
}
subreq->source = NETFS_DOWNLOAD_FROM_SERVER;
subreq->start = start;
subreq->len = len;
subreq->stream_nr = stream->stream_nr;
subreq->retry_count = 1;
trace_netfs_sreq_ref(rreq->debug_id, subreq->debug_index,
refcount_read(&subreq->ref),
netfs_sreq_trace_new);
list_add(&subreq->rreq_link, &to->rreq_link);
to = list_next_entry(to, rreq_link);
trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
stream->sreq_max_len = umin(len, rreq->rsize);
stream->sreq_max_segs = 0;
if (unlikely(stream->sreq_max_segs))
part = netfs_limit_iter(&source, 0, part, stream->sreq_max_segs);
netfs_stat(&netfs_n_rh_download);
if (rreq->netfs_ops->prepare_read(subreq) < 0) {
trace_netfs_sreq(subreq, netfs_sreq_trace_reprep_failed);
__set_bit(NETFS_SREQ_FAILED, &subreq->flags);
goto abandon;
}
part = umin(len, stream->sreq_max_len);
subreq->len = subreq->transferred + part;
subreq->io_iter = source;
iov_iter_truncate(&subreq->io_iter, part);
iov_iter_advance(&source, part);
len -= part;
start += part;
if (!len && boundary) {
__set_bit(NETFS_SREQ_BOUNDARY, &to->flags);
boundary = false;
}
netfs_reissue_read(rreq, subreq);
} while (len);
} while (!list_is_head(next, &stream->subrequests));
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
return;
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
/* If we hit an error, fail all remaining incomplete subrequests */
abandon_after:
if (list_is_last(&subreq->rreq_link, &stream->subrequests))
return;
subreq = list_next_entry(subreq, rreq_link);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
abandon:
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
list_for_each_entry_from(subreq, &stream->subrequests, rreq_link) {
if (!subreq->error &&
!test_bit(NETFS_SREQ_FAILED, &subreq->flags) &&
!test_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags))
continue;
subreq->error = -ENOMEM;
__set_bit(NETFS_SREQ_FAILED, &subreq->flags);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
__clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
}
}
/*
* Retry reads.
*/
void netfs_retry_reads(struct netfs_io_request *rreq)
{
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
struct netfs_io_subrequest *subreq;
struct netfs_io_stream *stream = &rreq->io_streams[0];
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
DEFINE_WAIT(myself);
netfs: Add retry stat counters Add stat counters to count the number of request and subrequest retries and display them in /proc/fs/netfs/stats. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-3-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:24:00 +00:00
netfs_stat(&netfs_n_rh_retry_read_req);
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
set_bit(NETFS_RREQ_RETRYING, &rreq->flags);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
/* Wait for all outstanding I/O to quiesce before performing retries as
* we may need to renegotiate the I/O sizes.
*/
list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
continue;
trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
for (;;) {
prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
break;
trace_netfs_sreq(subreq, netfs_sreq_trace_wait_for);
schedule();
trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
}
finish_wait(&rreq->waitq, &myself);
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
}
netfs: Fix a number of read-retry hangs Fix a number of hangs in the netfslib read-retry code, including: (1) netfs_reissue_read() doubles up the getting of references on subrequests, thereby leaking the subrequest and causing inode eviction to wait indefinitely. This can lead to the kernel reporting a hang in the filesystem's evict_inode(). Fix this by removing the get from netfs_reissue_read() and adding one to netfs_retry_read_subrequests() to deal with the one place that didn't double up. (2) The loop in netfs_retry_read_subrequests() that retries a sequence of failed subrequests doesn't record whether or not it retried the one that the "subreq" pointer points to when it leaves the loop. It may not if renegotiation/repreparation of the subrequests means that fewer subrequests are needed to span the cumulative range of the sequence. Because it doesn't record this, the piece of code that discards now-superfluous subrequests doesn't know whether it should discard the one "subreq" points to - and so it doesn't. Fix this by noting whether the last subreq it examines is superfluous and if it is, then getting rid of it and all subsequent subrequests. If that one one wasn't superfluous, then we would have tried to go round the previous loop again and so there can be no further unretried subrequests in the sequence. (3) netfs_retry_read_subrequests() gets yet an extra ref on any additional subrequests it has to get because it ran out of ones it could reuse to to renegotiation/repreparation shrinking the subrequests. Fix this by removing that extra ref. (4) In netfs_retry_reads(), it was using wait_on_bit() to wait for NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the sequence - but netfs_read_subreq_terminated() is now using a wait queue on the request instead and so this wait will never finish. Fix this by waiting on the wait queue instead. To make this work, a new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell the wake-up code to wake up the wait queue rather than requeuing the request's work item. Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is no longer used. (5) Whilst not strictly anything to do with the hang, netfs_retry_read_subrequests() was also doubly incrementing the subreq_counter and re-setting the debug index, leaving a gap in the trace. This is also fixed. One of these hangs was observed with 9p and with cifs. Others were forced by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read() was created to provide an changing pattern of maximum subrequest sizes: static int afs_prepare_read(struct netfs_io_subrequest *subreq) { struct netfs_io_request *rreq = subreq->rreq; if (!S_ISREG(subreq->rreq->inode->i_mode)) return 0; if (subreq->retry_count < 20) rreq->io_streams[0].sreq_max_len = umax(200, 2222 - subreq->retry_count * 40); else rreq->io_streams[0].sreq_max_len = 3333; return 0; } and pointed to by afs_req_ops. Then the following: struct netfs_io_subrequest *subreq = op->fetch.subreq; if (subreq->error == 0 && S_ISREG(subreq->rreq->inode->i_mode) && subreq->retry_count < 20) { subreq->transferred = subreq->already_done; __clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags); __set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags); afs_fetch_data_notify(op); return; } was inserted into afs_fetch_data_success() at the beginning and struct netfs_io_subrequest given an extra field, "already_done" that was set to the value in "subreq->transferred" by netfs_reissue_read(). When reading a 4K file, the subrequests would get gradually smaller, a new subrequest would be allocated around the 3rd retry and then eventually be rendered superfluous when the 20th retry was hit and the limit on the first subrequest was eased. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com Tested-by: Marc Dionne <marc.dionne@auristor.com> Tested-by: Steve French <stfrench@microsoft.com> cc: Ihor Solodrai <ihor.solodrai@pm.me> cc: Eric Van Hensbergen <ericvh@kernel.org> cc: Latchesar Ionkov <lucho@ionkov.net> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Christian Schoenebeck <linux_oss@crudebyte.com> cc: Paulo Alcantara <pc@manguebit.com> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs@lists.linux.dev cc: linux-cifs@vger.kernel.org cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2025-02-12 22:23:59 +00:00
clear_bit(NETFS_RREQ_RETRYING, &rreq->flags);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs: Change the read result collector to only use one work item Change the way netfslib collects read results to do all the collection for a particular read request using a single work item that walks along the subrequest queue as subrequests make progress or complete, unlocking folios progressively rather than doing the unlock in parallel as parallel requests come in. The code is remodelled to be more like the write-side code, though only using a single stream. This makes it more directly comparable and thus easier to duplicate fixes between the two sides. This has a number of advantages: (1) It's simpler. There doesn't need to be a complex donation mechanism to handle mismatches between the size and alignment of subrequests and folios. The collector unlocks folios as the subrequests covering each complete. (2) It should cause less scheduler overhead as there's a single work item in play unlocking pages in parallel when a read gets split up into a lot of subrequests instead of one per subrequest. Whilst the parallellism is nice in theory, in practice, the vast majority of loads are sequential reads of the whole file, so committing a bunch of threads to unlocking folios out of order doesn't help in those cases. (3) It should make it easier to implement content decryption. A folio cannot be decrypted until all the requests that contribute to it have completed - and, again, most loads are sequential and so, most of the time, we want to begin decryption sequentially (though it's great if the decryption can happen in parallel). There is a disadvantage in that we're losing the ability to decrypt and unlock things on an as-things-arrive basis which may affect some applications. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:41:17 +00:00
trace_netfs_rreq(rreq, netfs_rreq_trace_resubmit);
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
netfs_retry_read_subrequests(rreq);
}
/*
* Unlock any the pages that haven't been unlocked yet due to abandoned
* subrequests.
*/
void netfs_unlock_abandoned_read_pages(struct netfs_io_request *rreq)
{
struct folio_queue *p;
netfs: Abstract out a rolling folio buffer implementation A rolling buffer is a series of folios held in a list of folio_queues. New folios and folio_queue structs may be inserted at the head simultaneously with spent ones being removed from the tail without the need for locking. The rolling buffer includes an iov_iter and it has to be careful managing this as the list of folio_queues is extended such that an oops doesn't incurred because the iterator was pointing to the end of a folio_queue segment that got appended to and then removed. We need to use the mechanism twice, once for read and once for write, and, in future patches, we will use a second rolling buffer to handle bounce buffering for content encryption. Signed-off-by: David Howells <dhowells@redhat.com> Link: https://lore.kernel.org/r/20241216204124.3752367-6-dhowells@redhat.com cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-12-16 20:40:55 +00:00
for (p = rreq->buffer.tail; p; p = p->next) {
netfs: Speed up buffered reading Improve the efficiency of buffered reads in a number of ways: (1) Overhaul the algorithm in general so that it's a lot more compact and split the read submission code between buffered and unbuffered versions. The unbuffered version can be vastly simplified. (2) Read-result collection is handed off to a work queue rather than being done in the I/O thread. Multiple subrequests can be processes simultaneously. (3) When a subrequest is collected, any folios it fully spans are collected and "spare" data on either side is donated to either the previous or the next subrequest in the sequence. Notes: (*) Readahead expansion is massively slows down fio, presumably because it causes a load of extra allocations, both folio and xarray, up front before RPC requests can be transmitted. (*) RDMA with cifs does appear to work, both with SIW and RXE. (*) PG_private_2-based reading and copy-to-cache is split out into its own file and altered to use folio_queue. Note that the copy to the cache now creates a new write transaction against the cache and adds the folios to be copied into it. This allows it to use part of the writeback I/O code. Signed-off-by: David Howells <dhowells@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: netfs@lists.linux.dev cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2 Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-07-02 00:40:22 +01:00
for (int slot = 0; slot < folioq_count(p); slot++) {
struct folio *folio = folioq_folio(p, slot);
if (folio && !folioq_is_marked2(p, slot)) {
trace_netfs_folio(folio, netfs_folio_trace_abandon);
folio_unlock(folio);
}
}
}
}
Reference in a new issue Copy permalink