Re: gcc 4.5.1 / as 2.20.51.0.11 miscompiling drivers/char/i8k.c ?

[Andi Kleen]

> That said, changing the inline asm to just clobber one less register
> would be completely sufficient to make it work well with all gccs out there,
> just push/pop one of the register around the whole body.  I doubt calling
> out SMM BIOS is actually so performance critical that one push and one pop
> would ruin it.  Of course x86_64 version can stay as is, there are enough
> registers left...

Yes traditionally clobbering all registers has been dangerous
and it clearly can be done inside the asm too.

Here's a untested patch to do some manual push/pops too. Could someone with
the hardware please test it? (running a 32bit kernel)

-Andi

---

i8k: Clobber less registers

gcc doesn't like inline assembler statements that clobber nearly
all registers. Save a few registers manually on i386 to avoid this
problem.

Fix suggested by Jakub Jelinek

Signed-off-by: Andi Kleen <ak@xxxxxxxxxxxxxxx>

diff --git a/drivers/char/i8k.c b/drivers/char/i8k.c
index f0863be..a2da38b 100644
--- a/drivers/char/i8k.c
+++ b/drivers/char/i8k.c
@@ -146,7 +146,10 @@ static int i8k_smm(struct smm_regs *regs)
 		:    "a"(regs)
 		:    "%ebx", "%ecx", "%edx", "%esi", "%edi", "memory");
 #else
-	asm("pushl %%eax\n\t"
+	asm("pushl %%ebx\n\t"
+	    "pushl %%ecx\n\t"
+	    "pushl %%edx\n\t"
+	    "pushl %%eax\n\t"
 	    "movl 0(%%eax),%%edx\n\t"
 	    "push %%edx\n\t"
 	    "movl 4(%%eax),%%ebx\n\t"
@@ -167,10 +170,13 @@ static int i8k_smm(struct smm_regs *regs)
 	    "movl %%edx,0(%%eax)\n\t"
 	    "lahf\n\t"
 	    "shrl $8,%%eax\n\t"
-	    "andl $1,%%eax\n"
+	    "andl $1,%%eax\n\t"
+	    "popl %%edx\n\t"
+	    "popl %%ecx\n\t"
+	    "popl %%ebx\n"
 	    :"=a"(rc), "+m" (*regs)
 	    :    "a"(regs)
-	    :    "%ebx", "%ecx", "%edx", "%esi", "%edi", "memory");
+	    :    "%esi", "%edi", "memory");
 #endif
 	if (rc != 0 || (regs->eax & 0xffff) == 0xffff || regs->eax == eax)
 		return -EINVAL;


-- 
ak@xxxxxxxxxxxxxxx -- Speaking for myself only.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/