public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n
@ 2024-12-09  9:41 Ard Biesheuvel
  2024-12-10 10:26 ` [tip: x86/boot] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12 tip-bot2 for Ard Biesheuvel
  0 siblings, 1 reply; 2+ messages in thread
From: Ard Biesheuvel @ 2024-12-09  9:41 UTC (permalink / raw)
  To: linux-kernel
  Cc: x86, Ard Biesheuvel, Ingo Molnar, Borislav Petkov,
	kernel test robot

From: Ard Biesheuvel <ardb@kernel.org>

In __startup_64(), the bool 'la57' can only assume the 'true' value if
CONFIG_X86_5LEVEL is enabled in the build, and generally, the compiler
can make this inference at build time, and elide any references to the
symbol 'level4_kernel_pgt', which may be undefined if 'la57' is false.

As it turns out, GCC 12 gets this wrong sometimes, and gives up with a
build error

   ld: arch/x86/kernel/head64.o: in function `__startup_64':
   head64.c:(.head.text+0xbd): undefined reference to `level4_kernel_pgt'

even though the reference is in unreachable code. Fix this by
duplicating the IS_ENABLED(CONFIG_X86_5LEVEL) in the conditional that
tests the value of 'la57'.

Cc: Ingo Molnar <mingo@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202412060403.efD8Kgb7-lkp@intel.com/
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
 arch/x86/kernel/head64.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 54f9a8faf212..22c9ba305ac1 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -186,7 +186,7 @@ unsigned long __head __startup_64(unsigned long p2v_offset,
 	pgd = &RIP_REL_REF(early_top_pgt)->pgd;
 	pgd[pgd_index(__START_KERNEL_map)] += load_delta;
 
-	if (la57) {
+	if (IS_ENABLED(CONFIG_X86_5LEVEL) && la57) {
 		p4d = (p4dval_t *)&RIP_REL_REF(level4_kernel_pgt);
 		p4d[MAX_PTRS_PER_P4D - 1] += load_delta;
 
-- 
2.47.0.338.g60cca15819-goog


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [tip: x86/boot] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12
  2024-12-09  9:41 [PATCH] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n Ard Biesheuvel
@ 2024-12-10 10:26 ` tip-bot2 for Ard Biesheuvel
  0 siblings, 0 replies; 2+ messages in thread
From: tip-bot2 for Ard Biesheuvel @ 2024-12-10 10:26 UTC (permalink / raw)
  To: linux-tip-commits
  Cc: kernel test robot, Ard Biesheuvel, Ingo Molnar, x86, linux-kernel

The following commit has been merged into the x86/boot branch of tip:

Commit-ID:     35aafa1d41cee0d3d50164561bca34befc1d9ce3
Gitweb:        https://git.kernel.org/tip/35aafa1d41cee0d3d50164561bca34befc1d9ce3
Author:        Ard Biesheuvel <ardb@kernel.org>
AuthorDate:    Mon, 09 Dec 2024 10:41:06 +01:00
Committer:     Ingo Molnar <mingo@kernel.org>
CommitterDate: Tue, 10 Dec 2024 11:16:32 +01:00

x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12

In __startup_64(), the bool 'la57' can only assume the 'true' value if
CONFIG_X86_5LEVEL is enabled in the build, and generally, the compiler
can make this inference at build time, and elide any references to the
symbol 'level4_kernel_pgt', which may be undefined if 'la57' is false.

As it turns out, GCC 12 gets this wrong sometimes, and gives up with a
build error:

   ld: arch/x86/kernel/head64.o: in function `__startup_64':
   head64.c:(.head.text+0xbd): undefined reference to `level4_kernel_pgt'

even though the reference is in unreachable code. Fix this by
duplicating the IS_ENABLED(CONFIG_X86_5LEVEL) in the conditional that
tests the value of 'la57'.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20241209094105.762857-2-ardb+git@google.com
Closes: https://lore.kernel.org/oe-kbuild-all/202412060403.efD8Kgb7-lkp@intel.com/
---
 arch/x86/kernel/head64.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 54f9a8f..22c9ba3 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -186,7 +186,7 @@ unsigned long __head __startup_64(unsigned long p2v_offset,
 	pgd = &RIP_REL_REF(early_top_pgt)->pgd;
 	pgd[pgd_index(__START_KERNEL_map)] += load_delta;
 
-	if (la57) {
+	if (IS_ENABLED(CONFIG_X86_5LEVEL) && la57) {
 		p4d = (p4dval_t *)&RIP_REL_REF(level4_kernel_pgt);
 		p4d[MAX_PTRS_PER_P4D - 1] += load_delta;
 

^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2024-12-10 10:27 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-12-09  9:41 [PATCH] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n Ard Biesheuvel
2024-12-10 10:26 ` [tip: x86/boot] x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12 tip-bot2 for Ard Biesheuvel

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox