Skip to contentSkip to conversation

RAS: Fix inverted context info bounds check in ARM processor errors

PATCH Liuwenliang (Abbott Liu) 1 participant 1 message Sep 5 EDAC-core

lore
Share & tools
lore
#0 lore raw

Hi Hanjun, thinks for your review.

Quoted 13 lines

Hi Abbott,

Quoted 10 lines · Abbott Liu

Commit 87880af2d24e ("APEI/GHES: ARM processor Error: don't go past allocated memory") added bounds checks for malformed ARM processor error records but contained a bug:

In log_arm_hw_error(), the ctx_info bounds check is inverted. The condition sz + (long)ctx_info - (long)err >= err->section_length adds ctx_info->size when the context header is already past the end of the section instead of when it is within bounds. So change the comparison to <=.

Quoted 4 lines
Quoted 2 lines
Fixes 87880af2d24e ("APEI/GHES: ARM processor Error: don't go past allocated memory")

This empty line is not needed.

I am very sorry for making such a basic mistake; this issue will be resolved in the next version.

Quoted 22 lines
Quoted 11 lines
---
  drivers/ras/ras.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ras/ras.c +0 -0
index 03df3db62334..2540538a16a8 100644

drivers/ras/ras.c void log_arm_hw_error(struct cper_sec_proc_arm *err, const u8 sev) · line 78

Quoted 6 lines
drivers/ras/ras.c +1 -1
@@ -74,7 +74,7 @@ void log_arm_hw_error(struct cper_sec_proc_arm *err, const u8 sev)
  	for (n = 0; n < err->context_info_num; n++) {
  		sz = sizeof(struct cper_arm_ctx_info);
 
-		if (sz + (long)ctx_info - (long)err >= err->section_length)
+		if (sz + (long)ctx_info - (long)err <= err->section_length)
  			sz += ctx_info->size;

sz is an int and ctx_info->size is u32, if ctx_info->size is big enough for example over 0x7fffffff, the sz will be negative.

Quoted 1 line
Quoted 6 lines
Quoted 1 line
  		ctx_info = (struct cper_arm_ctx_info *)((long)ctx_info + sz);

if the sz is negative, the ctx_info may pointer to a wrong place.

ctx_info->size will not over 0x7fffffff in practical but should we consider the overflows?

I also think that it won't actually happen in practice, so there isn't much need to consider this overflow issue. But even if we do consider it, the code modifications wouldn't be too complex. I plan to fix it in the next version.

Quoted 6 lines

Thanks
Hanjun

Quoted 1 line