Skip to contentSkip to conversation

x86/mm: Check if PTRS_PER_PMD is defined before use

PATCH Superseded Andy Shevchenko 3 participants 5 messages Feb 24, 2025 x86

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/mm: Check if PTRS_PER_PMD is defined before use

Compiler is not happy about PTRS_PER_PMD being undefined

In file included from arch/x86/kernel/head_32.S:29:
arch/x86/include/asm/pgtable_32.h:59:5: error: "PTRS_PER_PMD" is not defined, evaluates to 0 [-Werror=undef]
59 | #if PTRS_PER_PMD > 1

Add a check to make sure PTRS_PER_PMD is defined before use.

---
 arch/x86/include/asm/pgtable_32.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/include/asm/pgtable_32.h +1 -1
index 7d4ad8907297..3c0523588f59 100644
@@ -56,7 +56,7 @@ do {						\
  * With PAE paging (PTRS_PER_PMD > 1), we allocate PTRS_PER_PGD == 4 pages for
  * the PMD's in addition to the pages required for the last level pagetables.
  */
-#if PTRS_PER_PMD > 1
+#if defined(PTRS_PER_PMD) && (PTRS_PER_PMD > 1)
 #define PAGE_TABLE_SIZE(pages) (((pages) / PTRS_PER_PMD) + PTRS_PER_PGD)
 #else
 #define PAGE_TABLE_SIZE(pages) ((pages) / PTRS_PER_PGD)
Signature · 1 line
2.45.1.3035.g276e886db78b
#1 lore raw
Quoted 7 lines · Andy Shevchenko

Compiler is not happy about PTRS_PER_PMD being undefined

In file included from arch/x86/kernel/head_32.S:29:
arch/x86/include/asm/pgtable_32.h:59:5: error: "PTRS_PER_PMD" is not defined, evaluates to 0 [-Werror=undef]
59 | #if PTRS_PER_PMD > 1

Add a check to make sure PTRS_PER_PMD is defined before use.

Hi Andy,

From reading the "Closes:" link, it appears this is a new issue that originates from a new compile flag. So it doesn't seem like it's worth backporting.

Also, the _behavior_ of "#if PTRS_PER_PMD > 1" was fine, right? It didn't cause the logic to go backwards from what was intended, does it?

This:

	https://gcc.gnu.org/onlinedocs/gcc-3.0.2/cpp_4.html

says: "Identifiers that are not macros, which are all considered to be the number zero."

Which would yield the correct behavior.

So I think this is purely a fix for new warning in new kernels. We shouldn't need to backport this anywhere at all.

#2 lore raw
Quoted 7 lines · Dave Hansen
Quoted 6 lines · Andy Shevchenko

Compiler is not happy about PTRS_PER_PMD being undefined

In file included from arch/x86/kernel/head_32.S:29:
arch/x86/include/asm/pgtable_32.h:59:5: error: "PTRS_PER_PMD" is not defined, evaluates to 0 [-Werror=undef]
59 | #if PTRS_PER_PMD > 1

Quoted 7 lines
Quoted 1 line

Add a check to make sure PTRS_PER_PMD is defined before use.

Hi Andy,

From reading the "Closes:" link, it appears this is a new issue that originates from a new compile flag. So it doesn't seem like it's worth backporting.

FWIW, I haven't put any Fixes tag nor Cc: stable@ :-) Also note this looks like both compilers complain about the same.

Quoted 14 lines

Also, the _behavior_ of "#if PTRS_PER_PMD > 1" was fine, right? It didn't cause the logic to go backwards from what was intended, does it?

This:

	https://gcc.gnu.org/onlinedocs/gcc-3.0.2/cpp_4.html

says: "Identifiers that are not macros, which are all considered to be the number zero."

Which would yield the correct behavior.

So I think this is purely a fix for new warning in new kernels. We shouldn't need to backport this anywhere at all.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#3 lore raw
Quoted 5 lines · Andy Shevchenko
Quoted 3 lines

From reading the "Closes:" link, it appears this is a new issue that originates from a new compile flag. So it doesn't seem like it's worth backporting.

FWIW, I haven't put any Fixes tag nor Cc: stable@ 🙂 Also note this looks like both compilers complain about the same.

Yeah, I did note that.

I see a lot of patches where folks simply forget about stable and Fixes. Unfortunately, I'm not super clairvoyant and when I see an ambiguous changelog without stable and Fixes, I can't tell if it was an oversight or intentional.

It would be great if contributors could be less ambiguous and save us the trouble of having to ask!

#4 lore raw

Mon, Feb 24, 2025 at 10:37:50AM -0800, Dave Hansen kirjoitti:

Quoted 7 lines · Andy Shevchenko
Quoted 5 lines
Quoted 3 lines

From reading the "Closes:" link, it appears this is a new issue that originates from a new compile flag. So it doesn't seem like it's worth backporting.

FWIW, I haven't put any Fixes tag nor Cc: stable@ 🙂 Also note this looks like both compilers complain about the same.

Quoted 6 lines

Yeah, I did note that.

I see a lot of patches where folks simply forget about stable and Fixes. Unfortunately, I'm not super clairvoyant and when I see an ambiguous changelog without stable and Fixes, I can't tell if it was an oversight or intentional.

In this case it was intentional.

Quoted 2 lines

It would be great if contributors could be less ambiguous and save us the trouble of having to ask!

Sure. For _this_ change as it's clear now, can we proceed for v6.15?

Signature · 2 lines
With Best Regards,
Andy Shevchenko