On Fri, May 30, 2025 at 07:11:58AM +0200, Greg Kroah-Hartman wrote:
On Thu, May 29, 2025 at 10:40:17AM -0700, Pawan Gupta wrote:
On Wed, May 28, 2025 at 09:49:40PM -0700, Richard Narron wrote:
On Tue, 27 May 2025, Pawan Gupta wrote:
On Tue, May 27, 2025 at 12:31:47PM -0700, Richard Narron wrote:
On Fri, 23 May 2025, Guenter Roeck wrote:
On 5/20/25 06:49, Greg Kroah-Hartman wrote: > This is the start of the stable review cycle for the 5.15.184 release. > There are 59 patches in this series, all will be posted as a response > to this one. If anyone has any issues with these being applied, please > let me know. > > Responses should be made by Thu, 22 May 2025 12:57:37 +0000. > Anything received after that time might be too late. >
Build reference: v5.15.184 Compiler version: x86_64-linux-gcc (GCC) 12.4.0 Assembler version: GNU assembler (GNU Binutils) 2.40
Configuration file workarounds: "s/CONFIG_FRAME_WARN=.*/CONFIG_FRAME_WARN=0/"
Building i386:defconfig ... passed Building i386:allyesconfig ... failed
Error log: x86_64-linux-ld: arch/x86/kernel/static_call.o: in function `__static_call_transform': static_call.c:(.ref.text+0x46): undefined reference to `cpu_wants_rethunk_at' make[1]: [Makefile:1234: vmlinux] Error 1 (ignored)
Building i386:allmodconfig ... failed
Error log: x86_64-linux-ld: arch/x86/kernel/static_call.o: in function `__static_call_transform': static_call.c:(.ref.text+0x46): undefined reference to `cpu_wants_rethunk_at' make[1]: [Makefile:1234: vmlinux] Error 1 (ignored)
In v5.15.y, cpu_wants_rethunk_at is only built if CONFIG_STACK_VALIDATION=y, but that is not supported for i386 builds. The dummy function in arch/x86/include/asm/alternative.h doesn't take that dependency into account.
I found this bug too using the Slackware 15.0 32-bit kernel configuration.
Here is a simple work around patch, but there may be a better solution...
--- arch/x86/kernel/static_call.c.orig 2025-05-22 05:08:28.000000000 -0700 +++ arch/x86/kernel/static_call.c 2025-05-27 10:25:27.630496538 -0700 @@ -81,9 +81,12 @@ break;
case RET:
+#ifdef CONFIG_64BIT if (cpu_wants_rethunk_at(insn)) code = text_gen_insn(JMP32_INSN_OPCODE, insn, x86_return_thunk); else +#endif code = &retinsn; break;
Another option is to define the empty function when CONFIG_STACK_VALIDATION=n as below:
--- 8< --- From: Pawan Gupta pawan.kumar.gupta@linux.intel.com Subject: [PATCH] x86/its: Fix undefined reference to cpu_wants_rethunk_at()
Below error was reported in 32-bit kernel build:
static_call.c:(.ref.text+0x46): undefined reference to `cpu_wants_rethunk_at' make[1]: [Makefile:1234: vmlinux] Error
This is because the definition of cpu_wants_rethunk_at() depends on CONFIG_STACK_VALIDATION which is only enabled in 64-bit mode.
Define the empty function when CONFIG_STACK_VALIDATION=n, rethunk mitigation is anyways not supported without it.
Reported-by: Guenter Roeck linux@roeck-us.net Link: https://lore.kernel.org/stable/0f597436-5da6-4319-b918-9f57bde5634a@roeck-us... Signed-off-by: Pawan Gupta pawan.kumar.gupta@linux.intel.com
arch/x86/include/asm/alternative.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index 1797f80c10de..a5f704dbb4a1 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -98,7 +98,7 @@ static inline u8 *its_static_thunk(int reg) } #endif
-#ifdef CONFIG_RETHUNK +#if defined(CONFIG_RETHUNK) && defined(CONFIG_STACK_VALIDATION) extern bool cpu_wants_rethunk(void); extern bool cpu_wants_rethunk_at(void *addr);
#else
2.34.1
Thanks for looking at this Pawan.
Your new patch works fine on both Slackware 15.0 32-bit and 64-bit systems.
Thank you for verifying the patch.
Is someone going to submit this in a form we can apply it in? (hint...)
Now submitted here:
https://lore.kernel.org/all/8c84125f71aec2fd81adf423dbc12156ac11706a.1748584...