linux/drivers/net/ethernet/sfc/siena/io.h
Baoquan He 0b1f77e74b asm-generic/iomap.h: remove ARCH_HAS_IOREMAP_xx macros
Patch series "mm: ioremap: Convert architectures to take GENERIC_IOREMAP
way", v8.

Motivation and implementation:
==============================
Currently, many architecutres have't taken the standard GENERIC_IOREMAP
way to implement ioremap_prot(), iounmap(), and ioremap_xx(), but make
these functions specifically under each arch's folder.  Those cause many
duplicated code of ioremap() and iounmap().

In this patchset, firstly introduce generic_ioremap_prot() and
generic_iounmap() to extract the generic code for GENERIC_IOREMAP.  By
taking GENERIC_IOREMAP method, the generic generic_ioremap_prot(),
generic_iounmap(), and their generic wrapper ioremap_prot(), ioremap() and
iounmap() are all visible and available to arch.  Arch needs to provide
wrapper functions to override the generic version if there's arch specific
handling in its corresponding ioremap_prot(), ioremap() or iounmap(). 
With these changes, duplicated ioremap/iounmap() code uder ARCH-es are
removed, and the equivalent functioality is kept as before.

Background info:
================

1) The converting more architectures to take GENERIC_IOREMAP way is
   suggested by Christoph in below discussion:
   https://lore.kernel.org/all/Yp7h0Jv6vpgt6xdZ@infradead.org/T/#u

2) In the previous v1 to v3, it's basically further action after arm64
   has converted to GENERIC_IOREMAP way in below patchset.  It's done by
   adding hook ioremap_allowed() and iounmap_allowed() in ARCH to add ARCH
   specific handling the middle of ioremap_prot() and iounmap().

[PATCH v5 0/6] arm64: Cleanup ioremap() and support ioremap_prot()
https://lore.kernel.org/all/20220607125027.44946-1-wangkefeng.wang@huawei.com/T/#u

Later, during v3 reviewing, Christophe Leroy suggested to introduce
generic_ioremap_prot() and generic_iounmap() to generic codes, and ARCH
can provide wrapper function ioremap_prot(), ioremap() or iounmap() if
needed.  Christophe made a RFC patchset as below to specially demonstrate
his idea.  This is what v4 and now v5 is doing.

[RFC PATCH 0/8] mm: ioremap: Convert architectures to take GENERIC_IOREMAP way
https://lore.kernel.org/all/cover.1665568707.git.christophe.leroy@csgroup.eu/T/#u

Testing:
========
In v8, I only applied this patchset onto the latest linus's tree to build
and run on arm64 and s390.


This patch (of 19):

Let's use '#define ioremap_xx' and "#ifdef ioremap_xx" instead.

To remove defined ARCH_HAS_IOREMAP_xx macros in <asm/io.h> of each ARCH,
the ARCH's own ioremap_wc|wt|np definition need be above "#include
<asm-generic/iomap.h>.  Otherwise the redefinition error would be seen
during compiling.  So the relevant adjustments are made to avoid compiling
error:

  loongarch:
  - doesn't include <asm-generic/iomap.h>, defining ARCH_HAS_IOREMAP_WC
    is redundant, so simply remove it.

  m68k:
  - selected GENERIC_IOMAP, <asm-generic/iomap.h> has been added in
    <asm-generic/io.h>, and <asm/kmap.h> is included above
    <asm-generic/iomap.h>, so simply remove ARCH_HAS_IOREMAP_WT defining.

  mips:
  - move "#include <asm-generic/iomap.h>" below ioremap_wc definition
    in <asm/io.h>

  powerpc:
  - remove "#include <asm-generic/iomap.h>" in <asm/io.h> because it's
    duplicated with the one in <asm-generic/io.h>, let's rely on the
    latter.

  x86:
  - selected GENERIC_IOMAP, remove #include <asm-generic/iomap.h> in
    the middle of <asm/io.h>. Let's rely on <asm-generic/io.h>.

Link: https://lkml.kernel.org/r/20230706154520.11257-2-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: David Laight <David.Laight@ACULAB.COM>
Cc: Helge Deller <deller@gmx.de>
Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Niklas Schnelle <schnelle@linux.ibm.com>
Cc: Stafford Horne <shorne@gmail.com>
Cc: Brian Cain <bcain@quicinc.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Rich Felker <dalias@libc.org>
Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Vineet Gupta <vgupta@kernel.org>
Cc: Will Deacon <will@kernel.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2023-08-18 10:12:32 -07:00

310 lines
9.9 KiB
C

/* SPDX-License-Identifier: GPL-2.0-only */
/****************************************************************************
* Driver for Solarflare network controllers and boards
* Copyright 2005-2006 Fen Systems Ltd.
* Copyright 2006-2013 Solarflare Communications Inc.
*/
#ifndef EFX_IO_H
#define EFX_IO_H
#include <linux/io.h>
#include <linux/spinlock.h>
/**************************************************************************
*
* NIC register I/O
*
**************************************************************************
*
* Notes on locking strategy for the Falcon architecture:
*
* Many CSRs are very wide and cannot be read or written atomically.
* Writes from the host are buffered by the Bus Interface Unit (BIU)
* up to 128 bits. Whenever the host writes part of such a register,
* the BIU collects the written value and does not write to the
* underlying register until all 4 dwords have been written. A
* similar buffering scheme applies to host access to the NIC's 64-bit
* SRAM.
*
* Writes to different CSRs and 64-bit SRAM words must be serialised,
* since interleaved access can result in lost writes. We use
* efx_nic::biu_lock for this.
*
* We also serialise reads from 128-bit CSRs and SRAM with the same
* spinlock. This may not be necessary, but it doesn't really matter
* as there are no such reads on the fast path.
*
* The DMA descriptor pointers (RX_DESC_UPD and TX_DESC_UPD) are
* 128-bit but are special-cased in the BIU to avoid the need for
* locking in the host:
*
* - They are write-only.
* - The semantics of writing to these registers are such that
* replacing the low 96 bits with zero does not affect functionality.
* - If the host writes to the last dword address of such a register
* (i.e. the high 32 bits) the underlying register will always be
* written. If the collector and the current write together do not
* provide values for all 128 bits of the register, the low 96 bits
* will be written as zero.
* - If the host writes to the address of any other part of such a
* register while the collector already holds values for some other
* register, the write is discarded and the collector maintains its
* current state.
*
* The EF10 architecture exposes very few registers to the host and
* most of them are only 32 bits wide. The only exceptions are the MC
* doorbell register pair, which has its own latching, and
* TX_DESC_UPD, which works in a similar way to the Falcon
* architecture.
*/
#if BITS_PER_LONG == 64
#define EFX_USE_QWORD_IO 1
#endif
/* Hardware issue requires that only 64-bit naturally aligned writes
* are seen by hardware. Its not strictly necessary to restrict to
* x86_64 arch, but done for safety since unusual write combining behaviour
* can break PIO.
*/
#ifdef CONFIG_X86_64
/* PIO is a win only if write-combining is possible */
#ifdef ioremap_wc
#define EFX_USE_PIO 1
#endif
#endif
static inline u32 efx_reg(struct efx_nic *efx, unsigned int reg)
{
return efx->reg_base + reg;
}
#ifdef EFX_USE_QWORD_IO
static inline void _efx_writeq(struct efx_nic *efx, __le64 value,
unsigned int reg)
{
__raw_writeq((__force u64)value, efx->membase + reg);
}
static inline __le64 _efx_readq(struct efx_nic *efx, unsigned int reg)
{
return (__force __le64)__raw_readq(efx->membase + reg);
}
#endif
static inline void _efx_writed(struct efx_nic *efx, __le32 value,
unsigned int reg)
{
__raw_writel((__force u32)value, efx->membase + reg);
}
static inline __le32 _efx_readd(struct efx_nic *efx, unsigned int reg)
{
return (__force __le32)__raw_readl(efx->membase + reg);
}
/* Write a normal 128-bit CSR, locking as appropriate. */
static inline void efx_writeo(struct efx_nic *efx, const efx_oword_t *value,
unsigned int reg)
{
unsigned long flags __attribute__ ((unused));
netif_vdbg(efx, hw, efx->net_dev,
"writing register %x with " EFX_OWORD_FMT "\n", reg,
EFX_OWORD_VAL(*value));
spin_lock_irqsave(&efx->biu_lock, flags);
#ifdef EFX_USE_QWORD_IO
_efx_writeq(efx, value->u64[0], reg + 0);
_efx_writeq(efx, value->u64[1], reg + 8);
#else
_efx_writed(efx, value->u32[0], reg + 0);
_efx_writed(efx, value->u32[1], reg + 4);
_efx_writed(efx, value->u32[2], reg + 8);
_efx_writed(efx, value->u32[3], reg + 12);
#endif
spin_unlock_irqrestore(&efx->biu_lock, flags);
}
/* Write 64-bit SRAM through the supplied mapping, locking as appropriate. */
static inline void efx_sram_writeq(struct efx_nic *efx, void __iomem *membase,
const efx_qword_t *value, unsigned int index)
{
unsigned int addr = index * sizeof(*value);
unsigned long flags __attribute__ ((unused));
netif_vdbg(efx, hw, efx->net_dev,
"writing SRAM address %x with " EFX_QWORD_FMT "\n",
addr, EFX_QWORD_VAL(*value));
spin_lock_irqsave(&efx->biu_lock, flags);
#ifdef EFX_USE_QWORD_IO
__raw_writeq((__force u64)value->u64[0], membase + addr);
#else
__raw_writel((__force u32)value->u32[0], membase + addr);
__raw_writel((__force u32)value->u32[1], membase + addr + 4);
#endif
spin_unlock_irqrestore(&efx->biu_lock, flags);
}
/* Write a 32-bit CSR or the last dword of a special 128-bit CSR */
static inline void efx_writed(struct efx_nic *efx, const efx_dword_t *value,
unsigned int reg)
{
netif_vdbg(efx, hw, efx->net_dev,
"writing register %x with "EFX_DWORD_FMT"\n",
reg, EFX_DWORD_VAL(*value));
/* No lock required */
_efx_writed(efx, value->u32[0], reg);
}
/* Read a 128-bit CSR, locking as appropriate. */
static inline void efx_reado(struct efx_nic *efx, efx_oword_t *value,
unsigned int reg)
{
unsigned long flags __attribute__ ((unused));
spin_lock_irqsave(&efx->biu_lock, flags);
value->u32[0] = _efx_readd(efx, reg + 0);
value->u32[1] = _efx_readd(efx, reg + 4);
value->u32[2] = _efx_readd(efx, reg + 8);
value->u32[3] = _efx_readd(efx, reg + 12);
spin_unlock_irqrestore(&efx->biu_lock, flags);
netif_vdbg(efx, hw, efx->net_dev,
"read from register %x, got " EFX_OWORD_FMT "\n", reg,
EFX_OWORD_VAL(*value));
}
/* Read 64-bit SRAM through the supplied mapping, locking as appropriate. */
static inline void efx_sram_readq(struct efx_nic *efx, void __iomem *membase,
efx_qword_t *value, unsigned int index)
{
unsigned int addr = index * sizeof(*value);
unsigned long flags __attribute__ ((unused));
spin_lock_irqsave(&efx->biu_lock, flags);
#ifdef EFX_USE_QWORD_IO
value->u64[0] = (__force __le64)__raw_readq(membase + addr);
#else
value->u32[0] = (__force __le32)__raw_readl(membase + addr);
value->u32[1] = (__force __le32)__raw_readl(membase + addr + 4);
#endif
spin_unlock_irqrestore(&efx->biu_lock, flags);
netif_vdbg(efx, hw, efx->net_dev,
"read from SRAM address %x, got "EFX_QWORD_FMT"\n",
addr, EFX_QWORD_VAL(*value));
}
/* Read a 32-bit CSR or SRAM */
static inline void efx_readd(struct efx_nic *efx, efx_dword_t *value,
unsigned int reg)
{
value->u32[0] = _efx_readd(efx, reg);
netif_vdbg(efx, hw, efx->net_dev,
"read from register %x, got "EFX_DWORD_FMT"\n",
reg, EFX_DWORD_VAL(*value));
}
/* Write a 128-bit CSR forming part of a table */
static inline void
efx_writeo_table(struct efx_nic *efx, const efx_oword_t *value,
unsigned int reg, unsigned int index)
{
efx_writeo(efx, value, reg + index * sizeof(efx_oword_t));
}
/* Read a 128-bit CSR forming part of a table */
static inline void efx_reado_table(struct efx_nic *efx, efx_oword_t *value,
unsigned int reg, unsigned int index)
{
efx_reado(efx, value, reg + index * sizeof(efx_oword_t));
}
/* default VI stride (step between per-VI registers) is 8K on EF10 and
* 64K on EF100
*/
#define EFX_DEFAULT_VI_STRIDE 0x2000
#define EF100_DEFAULT_VI_STRIDE 0x10000
/* Calculate offset to page-mapped register */
static inline unsigned int efx_paged_reg(struct efx_nic *efx, unsigned int page,
unsigned int reg)
{
return page * efx->vi_stride + reg;
}
/* Write the whole of RX_DESC_UPD or TX_DESC_UPD */
static inline void _efx_writeo_page(struct efx_nic *efx, efx_oword_t *value,
unsigned int reg, unsigned int page)
{
reg = efx_paged_reg(efx, page, reg);
netif_vdbg(efx, hw, efx->net_dev,
"writing register %x with " EFX_OWORD_FMT "\n", reg,
EFX_OWORD_VAL(*value));
#ifdef EFX_USE_QWORD_IO
_efx_writeq(efx, value->u64[0], reg + 0);
_efx_writeq(efx, value->u64[1], reg + 8);
#else
_efx_writed(efx, value->u32[0], reg + 0);
_efx_writed(efx, value->u32[1], reg + 4);
_efx_writed(efx, value->u32[2], reg + 8);
_efx_writed(efx, value->u32[3], reg + 12);
#endif
}
#define efx_writeo_page(efx, value, reg, page) \
_efx_writeo_page(efx, value, \
reg + \
BUILD_BUG_ON_ZERO((reg) != 0x830 && (reg) != 0xa10), \
page)
/* Write a page-mapped 32-bit CSR (EVQ_RPTR, EVQ_TMR (EF10), or the
* high bits of RX_DESC_UPD or TX_DESC_UPD)
*/
static inline void
_efx_writed_page(struct efx_nic *efx, const efx_dword_t *value,
unsigned int reg, unsigned int page)
{
efx_writed(efx, value, efx_paged_reg(efx, page, reg));
}
#define efx_writed_page(efx, value, reg, page) \
_efx_writed_page(efx, value, \
reg + \
BUILD_BUG_ON_ZERO((reg) != 0x180 && \
(reg) != 0x200 && \
(reg) != 0x400 && \
(reg) != 0x420 && \
(reg) != 0x830 && \
(reg) != 0x83c && \
(reg) != 0xa18 && \
(reg) != 0xa1c), \
page)
/* Write TIMER_COMMAND. This is a page-mapped 32-bit CSR, but a bug
* in the BIU means that writes to TIMER_COMMAND[0] invalidate the
* collector register.
*/
static inline void _efx_writed_page_locked(struct efx_nic *efx,
const efx_dword_t *value,
unsigned int reg,
unsigned int page)
{
unsigned long flags __attribute__ ((unused));
if (page == 0) {
spin_lock_irqsave(&efx->biu_lock, flags);
efx_writed(efx, value, efx_paged_reg(efx, page, reg));
spin_unlock_irqrestore(&efx->biu_lock, flags);
} else {
efx_writed(efx, value, efx_paged_reg(efx, page, reg));
}
}
#define efx_writed_page_locked(efx, value, reg, page) \
_efx_writed_page_locked(efx, value, \
reg + BUILD_BUG_ON_ZERO((reg) != 0x420), \
page)
#endif /* EFX_IO_H */