diff mbox series

[v5,01/17] asm-generic/iomap.h: remove ARCH_HAS_IOREMAP_xx macros

Message ID 20230301034247.136007-2-bhe@redhat.com
State New
Headers
Series mm: ioremap: Convert architectures to take GENERIC_IOREMAP way |

Commit Message

Baoquan He March 1, 2023, 3:42 a.m. UTC 
  Let's use '#define ioremap_xx' and "#ifdef ioremap_xx" instead.

For architectures defined ARCH_HAS_IOREMAP_xx macros in <asm/io.h>,
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, so <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> at
    the middle of <asm/io.h>. Let's rely on <asm-generic/io.h>.

This is a preparation patch so that  Later patch don't need to add
ARCH_HAS_IOREMAP_xx macro.

Signed-off-by: Baoquan He <bhe@redhat.com>
Cc: loongarch@lists.linux.dev
Cc: linux-m68k@lists.linux-m68k.org
Cc: linux-mips@vger.kernel.org
Cc: linuxppc-dev@lists.ozlabs.org
Cc: x86@kernel.org
Cc: netdev@vger.kernel.org
Cc: linux-arch@vger.kernel.org
---
 arch/loongarch/include/asm/io.h     | 2 --
 arch/m68k/include/asm/io_mm.h       | 2 --
 arch/m68k/include/asm/kmap.h        | 2 --
 arch/mips/include/asm/io.h          | 5 ++---
 arch/powerpc/include/asm/io.h       | 9 +--------
 arch/x86/include/asm/io.h           | 5 -----
 drivers/net/ethernet/sfc/io.h       | 2 +-
 drivers/net/ethernet/sfc/siena/io.h | 2 +-
 include/asm-generic/iomap.h         | 6 +++---
 9 files changed, 8 insertions(+), 27 deletions(-)

Comments

Edward Cree March 1, 2023, 4:38 a.m. UTC | #1 
On 01/03/2023 03:42, Baoquan He wrote:
> diff --git a/drivers/net/ethernet/sfc/io.h b/drivers/net/ethernet/sfc/io.h
> index 30439cc83a89..07f99ad14bf3 100644
> --- a/drivers/net/ethernet/sfc/io.h
> +++ b/drivers/net/ethernet/sfc/io.h
> @@ -70,7 +70,7 @@
>   */
>  #ifdef CONFIG_X86_64
>  /* PIO is a win only if write-combining is possible */
> -#ifdef ARCH_HAS_IOREMAP_WC
> +#ifdef ioremap_wc
>  #define EFX_USE_PIO 1
>  #endif
>  #endif

So I don't know how valid what we're doing here is...

> diff --git a/include/asm-generic/iomap.h b/include/asm-generic/iomap.h
> index 08237ae8b840..196087a8126e 100644
> --- a/include/asm-generic/iomap.h
> +++ b/include/asm-generic/iomap.h
> @@ -93,15 +93,15 @@ extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
>  extern void ioport_unmap(void __iomem *);
>  #endif
>  
> -#ifndef ARCH_HAS_IOREMAP_WC
> +#ifndef ioremap_wc
>  #define ioremap_wc ioremap
>  #endif

... but it looks like this will break it, since in sfc/io.h
 `#ifdef ioremap_wc` will always be true (if I'm correctly
 understanding what we get via #include <linux/io.h>, which I'm
 probably not because asm includes always confuse me).
I.e. we're not just interested in "can code that calls ioremap_wc
 compile?", we care about whether we actually get WC, because
 we're making an optimisation decision based on it.
Matthew Wilcox March 1, 2023, 5:11 a.m. UTC | #2 
On Wed, Mar 01, 2023 at 04:38:10AM +0000, Edward Cree wrote:
> On 01/03/2023 03:42, Baoquan He wrote:
> > diff --git a/drivers/net/ethernet/sfc/io.h b/drivers/net/ethernet/sfc/io.h
> > index 30439cc83a89..07f99ad14bf3 100644
> > --- a/drivers/net/ethernet/sfc/io.h
> > +++ b/drivers/net/ethernet/sfc/io.h
> > @@ -70,7 +70,7 @@
> >   */
> >  #ifdef CONFIG_X86_64
> >  /* PIO is a win only if write-combining is possible */
> > -#ifdef ARCH_HAS_IOREMAP_WC
> > +#ifdef ioremap_wc
> >  #define EFX_USE_PIO 1
> >  #endif
> >  #endif
> 
> So I don't know how valid what we're doing here is...

Well, x86 defines ARCH_HAS_IOREMAP_WC unconditionally, so it doesn't
affect you ... but you raise a good question about how a driver can
determine if it's actually getting WC memory.
Baoquan He March 1, 2023, 9:24 a.m. UTC | #3 
On 03/01/23 at 05:11am, Matthew Wilcox wrote:
> On Wed, Mar 01, 2023 at 04:38:10AM +0000, Edward Cree wrote:
> > On 01/03/2023 03:42, Baoquan He wrote:
> > > diff --git a/drivers/net/ethernet/sfc/io.h b/drivers/net/ethernet/sfc/io.h
> > > index 30439cc83a89..07f99ad14bf3 100644
> > > --- a/drivers/net/ethernet/sfc/io.h
> > > +++ b/drivers/net/ethernet/sfc/io.h
> > > @@ -70,7 +70,7 @@
> > >   */
> > >  #ifdef CONFIG_X86_64
> > >  /* PIO is a win only if write-combining is possible */
> > > -#ifdef ARCH_HAS_IOREMAP_WC
> > > +#ifdef ioremap_wc
> > >  #define EFX_USE_PIO 1
> > >  #endif
> > >  #endif
> > 
> > So I don't know how valid what we're doing here is...
> 
> Well, x86 defines ARCH_HAS_IOREMAP_WC unconditionally, so it doesn't
> affect you ... but you raise a good question about how a driver can
> determine if it's actually getting WC memory.

Yeah, this change doesn't affect sfc. Because ARCH_HAS_IOREMAP_WC is used to
make ioremap_wc defined in <asm/io.h> override the default one in
<asm-generic/iomap.h>, this patch has made code have the same effect.

Besides, I have a question still in my mind. Surely this is unrelated to
this patch.

In commit 38d9029a652c (parisc: Define ioremap_uc and ioremap_wc),
ioremap_wc definition was added in arch/parisc/include/asm/io.h, and it
didn't add ARCH_HAS_IOREMAP_WC definition. However, it won't cause
redefinition of ioremap_wc, even though there's "#include <asm-generic/iomap.h>"
at below. I could be dizzy on these io.h and iomap.h.

When I added ioremap_wt and ioremap_np to debug, ioremap_np will
cause redefinition, while ioremap_wt woundn't. Does anyone know what
I am missing?

diff --git a/arch/parisc/include/asm/io.h b/arch/parisc/include/asm/io.h
index c05e781be2f5..20d566eec3b3 100644
--- a/arch/parisc/include/asm/io.h
+++ b/arch/parisc/include/asm/io.h
@@ -127,6 +127,8 @@ static inline void gsc_writeq(unsigned long long val, unsigned long addr)
  */
 void __iomem *ioremap(unsigned long offset, unsigned long size);
 #define ioremap_wc                     ioremap
+#define ioremap_wt                     ioremap
+#define ioremap_np                     ioremap
 #define ioremap_uc                     ioremap
 #define pci_iounmap                    pci_iounmap
diff mbox series

Patch

diff --git a/arch/loongarch/include/asm/io.h b/arch/loongarch/include/asm/io.h
index 402a7d9e3a53..505e3b583463 100644
--- a/arch/loongarch/include/asm/io.h
+++ b/arch/loongarch/include/asm/io.h
@@ -5,8 +5,6 @@ 
 #ifndef _ASM_IO_H
 #define _ASM_IO_H
 
-#define ARCH_HAS_IOREMAP_WC
-
 #include <linux/kernel.h>
 #include <linux/types.h>
 
diff --git a/arch/m68k/include/asm/io_mm.h b/arch/m68k/include/asm/io_mm.h
index d41fa488453b..6a0abd4846c6 100644
--- a/arch/m68k/include/asm/io_mm.h
+++ b/arch/m68k/include/asm/io_mm.h
@@ -26,8 +26,6 @@ 
 #include <asm/virtconvert.h>
 #include <asm/kmap.h>
 
-#include <asm-generic/iomap.h>
-
 #ifdef CONFIG_ATARI
 #define atari_readb   raw_inb
 #define atari_writeb  raw_outb
diff --git a/arch/m68k/include/asm/kmap.h b/arch/m68k/include/asm/kmap.h
index dec05743d426..4efb3efa593a 100644
--- a/arch/m68k/include/asm/kmap.h
+++ b/arch/m68k/include/asm/kmap.h
@@ -4,8 +4,6 @@ 
 
 #ifdef CONFIG_MMU
 
-#define ARCH_HAS_IOREMAP_WT
-
 /* Values for nocacheflag and cmode */
 #define IOMAP_FULL_CACHING		0
 #define IOMAP_NOCACHE_SER		1
diff --git a/arch/mips/include/asm/io.h b/arch/mips/include/asm/io.h
index e6d5ccaa309e..cec8347f0b85 100644
--- a/arch/mips/include/asm/io.h
+++ b/arch/mips/include/asm/io.h
@@ -12,8 +12,6 @@ 
 #ifndef _ASM_IO_H
 #define _ASM_IO_H
 
-#define ARCH_HAS_IOREMAP_WC
-
 #include <linux/compiler.h>
 #include <linux/kernel.h>
 #include <linux/types.h>
@@ -25,7 +23,6 @@ 
 #include <asm/byteorder.h>
 #include <asm/cpu.h>
 #include <asm/cpu-features.h>
-#include <asm-generic/iomap.h>
 #include <asm/page.h>
 #include <asm/pgtable-bits.h>
 #include <asm/processor.h>
@@ -210,6 +207,8 @@  void iounmap(const volatile void __iomem *addr);
 #define ioremap_wc(offset, size)					\
 	ioremap_prot((offset), (size), boot_cpu_data.writecombine)
 
+#include <asm-generic/iomap.h>
+
 #if defined(CONFIG_CPU_CAVIUM_OCTEON) || defined(CONFIG_CPU_LOONGSON64)
 #define war_io_reorder_wmb()		wmb()
 #else
diff --git a/arch/powerpc/include/asm/io.h b/arch/powerpc/include/asm/io.h
index fc112a91d0c2..7a82e6f70ced 100644
--- a/arch/powerpc/include/asm/io.h
+++ b/arch/powerpc/include/asm/io.h
@@ -3,11 +3,6 @@ 
 #define _ASM_POWERPC_IO_H
 #ifdef __KERNEL__
 
-#define ARCH_HAS_IOREMAP_WC
-#ifdef CONFIG_PPC32
-#define ARCH_HAS_IOREMAP_WT
-#endif
-
 /*
  */
 
@@ -695,9 +690,7 @@  static inline void name at					\
 #define writel_relaxed(v, addr)	writel(v, addr)
 #define writeq_relaxed(v, addr)	writeq(v, addr)
 
-#ifdef CONFIG_GENERIC_IOMAP
-#include <asm-generic/iomap.h>
-#else
+#ifndef CONFIG_GENERIC_IOMAP
 /*
  * Here comes the implementation of the IOMAP interfaces.
  */
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index e9025640f634..76238842406a 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -35,9 +35,6 @@ 
   *  - Arnaldo Carvalho de Melo <acme@conectiva.com.br>
   */
 
-#define ARCH_HAS_IOREMAP_WC
-#define ARCH_HAS_IOREMAP_WT
-
 #include <linux/string.h>
 #include <linux/compiler.h>
 #include <linux/cc_platform.h>
@@ -212,8 +209,6 @@  void memset_io(volatile void __iomem *, int, size_t);
 #define memcpy_toio memcpy_toio
 #define memset_io memset_io
 
-#include <asm-generic/iomap.h>
-
 /*
  * ISA space is 'always mapped' on a typical x86 system, no need to
  * explicitly ioremap() it. The fact that the ISA IO space is mapped
diff --git a/drivers/net/ethernet/sfc/io.h b/drivers/net/ethernet/sfc/io.h
index 30439cc83a89..07f99ad14bf3 100644
--- a/drivers/net/ethernet/sfc/io.h
+++ b/drivers/net/ethernet/sfc/io.h
@@ -70,7 +70,7 @@ 
  */
 #ifdef CONFIG_X86_64
 /* PIO is a win only if write-combining is possible */
-#ifdef ARCH_HAS_IOREMAP_WC
+#ifdef ioremap_wc
 #define EFX_USE_PIO 1
 #endif
 #endif
diff --git a/drivers/net/ethernet/sfc/siena/io.h b/drivers/net/ethernet/sfc/siena/io.h
index 30439cc83a89..07f99ad14bf3 100644
--- a/drivers/net/ethernet/sfc/siena/io.h
+++ b/drivers/net/ethernet/sfc/siena/io.h
@@ -70,7 +70,7 @@ 
  */
 #ifdef CONFIG_X86_64
 /* PIO is a win only if write-combining is possible */
-#ifdef ARCH_HAS_IOREMAP_WC
+#ifdef ioremap_wc
 #define EFX_USE_PIO 1
 #endif
 #endif
diff --git a/include/asm-generic/iomap.h b/include/asm-generic/iomap.h
index 08237ae8b840..196087a8126e 100644
--- a/include/asm-generic/iomap.h
+++ b/include/asm-generic/iomap.h
@@ -93,15 +93,15 @@  extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
 extern void ioport_unmap(void __iomem *);
 #endif
 
-#ifndef ARCH_HAS_IOREMAP_WC
+#ifndef ioremap_wc
 #define ioremap_wc ioremap
 #endif
 
-#ifndef ARCH_HAS_IOREMAP_WT
+#ifndef ioremap_wt
 #define ioremap_wt ioremap
 #endif
 
-#ifndef ARCH_HAS_IOREMAP_NP
+#ifndef ioremap_np
 /* See the comment in asm-generic/io.h about ioremap_np(). */
 #define ioremap_np ioremap_np
 static inline void __iomem *ioremap_np(phys_addr_t offset, size_t size)