mirror of
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
synced 2025-08-04 16:25:34 +00:00
005b2a9dc8
-----BEGIN PGP SIGNATURE----- iQJEBAABCAAuFiEEwPw5LcreJtl1+l5K99NY+ylx4KYFAl/YJxsQHGF4Ym9lQGtl cm5lbC5kawAKCRD301j7KXHgpjpyEACBdW+YjenjTbkUPeEXzQgkBkTZUYw3g007 DPcUT1g8PQZXYXlQvBKCvGhhIr7/KVcjepKoowiNQfBNGcIPJTVopW58nzpqAfTQ goI2WYGn5EKFFKBPvtH04cJD/Wo8muXdxynKtqyZbnGGgZjQxPrE259b8dpHjBSR 6L7HHkk0D1oU/5b6h6Ocpg9mc/0iIUCZylySAYY3eGO0JaVPJaXgZSJZYgHxCHll Lb+/y/fXdtm/0PmQ3ko0ev54g3yEWqZIX0NsZW1asrButIy+KLzQ2Mz1xFLFDMag prtIfwb8tzgc4dFPY090C/azjCh5CPpxqYS6FkRwS0p86n6OhkyXrqfily5Hs4/B NC7CBPBSH/j+NKUK7CYZcpTzTpxPjUr9p0anUdlvMJz8FhTb/3YEEZ1UTeWOeHmk Yo5SxnFghLeZZeZ1ok6rdymnVa7WEX12SCLGQX31BB2mld0tNbKb4b+FsBF6OUMk IUaX6OjwDFVRaysC88BQ4hjcIP1HxsViG4/VZDX15gjAAH2Pvb+7tev+lcDcOhjz TCD4GNFspTFzRhh9nT7oxQ679qCh9G9zHbzuIRewnrS6iqvo5SJQB3dR2yrWZRRH ySkQFiHpYOlnLJYv0jg9COlGwo2FUdcvKhCvkjQKKBz48rzW/IC0LwKdRQWZDFk3 FKGzP/NBig== =cadT -----END PGP SIGNATURE----- Merge tag 'tif-task_work.arch-2020-12-14' of git://git.kernel.dk/linux-block Pull TIF_NOTIFY_SIGNAL updates from Jens Axboe: "This sits on top of of the core entry/exit and x86 entry branch from the tip tree, which contains the generic and x86 parts of this work. Here we convert the rest of the archs to support TIF_NOTIFY_SIGNAL. With that done, we can get rid of JOBCTL_TASK_WORK from task_work and signal.c, and also remove a deadlock work-around in io_uring around knowing that signal based task_work waking is invoked with the sighand wait queue head lock. The motivation for this work is to decouple signal notify based task_work, of which io_uring is a heavy user of, from sighand. The sighand lock becomes a huge contention point, particularly for threaded workloads where it's shared between threads. Even outside of threaded applications it's slower than it needs to be. Roman Gershman <romger@amazon.com> reported that his networked workload dropped from 1.6M QPS at 80% CPU to 1.0M QPS at 100% CPU after io_uring was changed to use TIF_NOTIFY_SIGNAL. The time was all spent hammering on the sighand lock, showing 57% of the CPU time there [1]. There are further cleanups possible on top of this. One example is TIF_PATCH_PENDING, where a patch already exists to use TIF_NOTIFY_SIGNAL instead. Hopefully this will also lead to more consolidation, but the work stands on its own as well" [1] https://github.com/axboe/liburing/issues/215 * tag 'tif-task_work.arch-2020-12-14' of git://git.kernel.dk/linux-block: (28 commits) io_uring: remove 'twa_signal_ok' deadlock work-around kernel: remove checking for TIF_NOTIFY_SIGNAL signal: kill JOBCTL_TASK_WORK io_uring: JOBCTL_TASK_WORK is no longer used by task_work task_work: remove legacy TWA_SIGNAL path sparc: add support for TIF_NOTIFY_SIGNAL riscv: add support for TIF_NOTIFY_SIGNAL nds32: add support for TIF_NOTIFY_SIGNAL ia64: add support for TIF_NOTIFY_SIGNAL h8300: add support for TIF_NOTIFY_SIGNAL c6x: add support for TIF_NOTIFY_SIGNAL alpha: add support for TIF_NOTIFY_SIGNAL xtensa: add support for TIF_NOTIFY_SIGNAL arm: add support for TIF_NOTIFY_SIGNAL microblaze: add support for TIF_NOTIFY_SIGNAL hexagon: add support for TIF_NOTIFY_SIGNAL csky: add support for TIF_NOTIFY_SIGNAL openrisc: add support for TIF_NOTIFY_SIGNAL sh: add support for TIF_NOTIFY_SIGNAL um: add support for TIF_NOTIFY_SIGNAL ... |
||
|---|---|---|
| .. | ||
| fpu | ||
| trace | ||
| vdso | ||
| airq.h | ||
| alternative-asm.h | ||
| alternative.h | ||
| ap.h | ||
| appldata.h | ||
| archrandom.h | ||
| asm-const.h | ||
| asm-prototypes.h | ||
| atomic.h | ||
| atomic_ops.h | ||
| barrier.h | ||
| bitops.h | ||
| boot_data.h | ||
| bug.h | ||
| bugs.h | ||
| cache.h | ||
| ccwdev.h | ||
| ccwgroup.h | ||
| checksum.h | ||
| chpid.h | ||
| chsc.h | ||
| cio.h | ||
| clocksource.h | ||
| clp.h | ||
| cmb.h | ||
| cmpxchg.h | ||
| compat.h | ||
| cpacf.h | ||
| cpcmd.h | ||
| cpu.h | ||
| cpu_mcf.h | ||
| cpu_mf-insn.h | ||
| cpu_mf.h | ||
| cpufeature.h | ||
| cputime.h | ||
| crw.h | ||
| css_chars.h | ||
| ctl_reg.h | ||
| current.h | ||
| debug.h | ||
| delay.h | ||
| diag.h | ||
| dis.h | ||
| dma.h | ||
| dwarf.h | ||
| eadm.h | ||
| ebcdic.h | ||
| elf.h | ||
| exec.h | ||
| extable.h | ||
| extmem.h | ||
| facility.h | ||
| fcx.h | ||
| ftrace.h | ||
| futex.h | ||
| gmap.h | ||
| hardirq.h | ||
| hugetlb.h | ||
| hw_irq.h | ||
| idals.h | ||
| idle.h | ||
| io.h | ||
| ipl.h | ||
| irq.h | ||
| irqflags.h | ||
| isc.h | ||
| itcw.h | ||
| jump_label.h | ||
| kasan.h | ||
| Kbuild | ||
| kdebug.h | ||
| kexec.h | ||
| kprobes.h | ||
| kvm_host.h | ||
| kvm_para.h | ||
| linkage.h | ||
| livepatch.h | ||
| lowcore.h | ||
| mem_detect.h | ||
| mem_encrypt.h | ||
| mmu.h | ||
| mmu_context.h | ||
| mmzone.h | ||
| module.h | ||
| nmi.h | ||
| nospec-branch.h | ||
| nospec-insn.h | ||
| numa.h | ||
| os_info.h | ||
| page-states.h | ||
| page.h | ||
| pci.h | ||
| pci_clp.h | ||
| pci_debug.h | ||
| pci_dma.h | ||
| pci_insn.h | ||
| pci_io.h | ||
| percpu.h | ||
| perf_event.h | ||
| pgalloc.h | ||
| pgtable.h | ||
| pkey.h | ||
| pnet.h | ||
| preempt.h | ||
| processor.h | ||
| ptdump.h | ||
| ptrace.h | ||
| purgatory.h | ||
| qdio.h | ||
| runtime_instr.h | ||
| schid.h | ||
| sclp.h | ||
| scsw.h | ||
| seccomp.h | ||
| sections.h | ||
| serial.h | ||
| set_memory.h | ||
| setup.h | ||
| shmparam.h | ||
| signal.h | ||
| sigp.h | ||
| smp.h | ||
| sparsemem.h | ||
| spinlock.h | ||
| spinlock_types.h | ||
| stacktrace.h | ||
| stp.h | ||
| string.h | ||
| switch_to.h | ||
| syscall.h | ||
| syscall_wrapper.h | ||
| sysinfo.h | ||
| termios.h | ||
| thread_info.h | ||
| timex.h | ||
| tlb.h | ||
| tlbflush.h | ||
| topology.h | ||
| uaccess.h | ||
| unistd.h | ||
| unwind.h | ||
| uprobes.h | ||
| user.h | ||
| uv.h | ||
| vdso.h | ||
| vga.h | ||
| vmalloc.h | ||
| vmlinux.lds.h | ||
| vtime.h | ||
| vtimer.h | ||
| vx-insn.h | ||
| xor.h | ||