Hi all,

After merging the origin tree, today's linux-next build (i386 defconfig)
failed like this:

x86_64-linux-gnu-ld: arch/x86/kernel/static_call.o: in function `__static_call_update_early':
static_call.c:(.noinstr.text+0x15): undefined reference to `static_call_initialized'

In case it matters, this is a PowerPC hosted cross build.

Presumably caused by commit

  0ef8047b737d ("x86/static-call: provide a way to do very early static-call updates")

static_call_initialized is defined in kernel/static_call_inline.c which is
only built if HAVE_STATIC_CALL_INLINE is defined and that is only selected
if HAVE_OBJTOOL is defined and that is only selected if X86_64 is defined.

I have applied the following hack for today - probably something better
can be done.

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Wed, 18 Dec 2024 15:05:03 +1100
Subject: [PATCH] fix up for "x86/static-call: provide a way to do very early
 static-call updates"

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 arch/x86/kernel/static_call.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/x86/kernel/static_call.c b/arch/x86/kernel/static_call.c
index 9eed0c144dad..dca440b18b9f 100644
--- a/arch/x86/kernel/static_call.c
+++ b/arch/x86/kernel/static_call.c
@@ -176,7 +176,9 @@ noinstr void __static_call_update_early(void *tramp, void *func)
 {
 	BUG_ON(system_state != SYSTEM_BOOTING);
 	BUG_ON(!early_boot_irqs_disabled);
+#ifdef CONFIG_HAVE_STATIC_CALL_INLINE
 	BUG_ON(static_call_initialized);
+#endif
 	__text_gen_insn(tramp, JMP32_INSN_OPCODE, tramp, func, JMP32_INSN_SIZE);
 	sync_core();
 }
-- 
2.45.2

-- 
Cheers,
Stephen Rothwell

On 18.12.24 05:13, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the origin tree, today's linux-next build (i386 defconfig)
> failed like this:
> 
> x86_64-linux-gnu-ld: arch/x86/kernel/static_call.o: in function `__static_call_update_early':
> static_call.c:(.noinstr.text+0x15): undefined reference to `static_call_initialized'
> 
> In case it matters, this is a PowerPC hosted cross build.
> 
> Presumably caused by commit
> 
>    0ef8047b737d ("x86/static-call: provide a way to do very early static-call updates")
> 
> static_call_initialized is defined in kernel/static_call_inline.c which is
> only built if HAVE_STATIC_CALL_INLINE is defined and that is only selected
> if HAVE_OBJTOOL is defined and that is only selected if X86_64 is defined.
> 
> I have applied the following hack for today - probably something better
> can be done.

Yeah, it doesn't cover all cases. I have sent a patch fixing it properly
(famous last words).


Juergen

> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Wed, 18 Dec 2024 15:05:03 +1100
> Subject: [PATCH] fix up for "x86/static-call: provide a way to do very early
>   static-call updates"
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>   arch/x86/kernel/static_call.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/arch/x86/kernel/static_call.c b/arch/x86/kernel/static_call.c
> index 9eed0c144dad..dca440b18b9f 100644
> --- a/arch/x86/kernel/static_call.c
> +++ b/arch/x86/kernel/static_call.c
> @@ -176,7 +176,9 @@ noinstr void __static_call_update_early(void *tramp, void *func)
>   {
>   	BUG_ON(system_state != SYSTEM_BOOTING);
>   	BUG_ON(!early_boot_irqs_disabled);
> +#ifdef CONFIG_HAVE_STATIC_CALL_INLINE
>   	BUG_ON(static_call_initialized);
> +#endif
>   	__text_gen_insn(tramp, JMP32_INSN_OPCODE, tramp, func, JMP32_INSN_SIZE);
>   	sync_core();
>   }