[PATCH] Before adding to usablemem_rgns, check if the memory range is already included.

[PATCH] Before adding to usablemem_rgns, check if the memory range is already included.

[Chen Haixiang]

When kexec_iomem_for_each_line() reads from /proc/iomem,
concurrent modifications to /proc/iomem may lead to
usablemem_rgns recording duplicate Crash kernel segments.
This can result in the number of retrieved Crash kernel
segments exceeding CRASH_MAX_RESERVED_RANGES, triggering
a realloc of the crash_reserved_mem in usablemem_rgns,
which could crash the process. We should ensure that each
range added to usablemem_rgns is unique to prevent these issues.
---
 kexec/arch/arm64/crashdump-arm64.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/kexec/arch/arm64/crashdump-arm64.c b/kexec/arch/arm64/crashdump-arm64.c
index 3098315..f23b2bf 100644
--- a/kexec/arch/arm64/crashdump-arm64.c
+++ b/kexec/arch/arm64/crashdump-arm64.c
@@ -63,10 +63,22 @@ static int iomem_range_callback(void *UNUSED(data), int UNUSED(nr),
 				char *str, unsigned long long base,
 				unsigned long long length)
 {
-	if (strncmp(str, CRASH_KERNEL, strlen(CRASH_KERNEL)) == 0)
+	int i;
+
+	if (strncmp(str, CRASH_KERNEL, strlen(CRASH_KERNEL)) == 0) {
+		/*
+		 * Checks whether the area exists in crash_reserved_mem.
+		 */
+		for (i = 0; i < usablemem_rgns.max_size; i++) {
+			if (usablemem_rgns.ranges[i].start == base) {
+				fprintf(stderr, "Warning, the range already exists in usablemem_rgns, base=%lx, length=%lx\n",
+						base, length);
+				return 0;
+			}
+		}
 		return mem_regions_alloc_and_add(&usablemem_rgns,
 						base, length, RANGE_RAM);
-	else if (strncmp(str, SYSTEM_RAM, strlen(SYSTEM_RAM)) == 0)
+	} else if (strncmp(str, SYSTEM_RAM, strlen(SYSTEM_RAM)) == 0)
 		return mem_regions_alloc_and_add(&system_memory_rgns,
 						base, length, RANGE_RAM);
 	else if (strncmp(str, KERNEL_CODE, strlen(KERNEL_CODE)) == 0) {
-- 
2.41.0


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH] Before adding to usablemem_rgns, check if the memory range is already included.

[Simon Horman]

On Thu, Oct 17, 2024 at 08:38:59PM +0800, Chen Haixiang wrote:
> When kexec_iomem_for_each_line() reads from /proc/iomem,
> concurrent modifications to /proc/iomem may lead to
> usablemem_rgns recording duplicate Crash kernel segments.
> This can result in the number of retrieved Crash kernel
> segments exceeding CRASH_MAX_RESERVED_RANGES, triggering
> a realloc of the crash_reserved_mem in usablemem_rgns,
> which could crash the process. We should ensure that each
> range added to usablemem_rgns is unique to prevent these issues.

Thanks Chen Haixiang,

This looks reasonable to me, but, as per Kernel submissions,
I will need an Signed-off-by line [1] in order to apply this patch.

[1] https://www.kernel.org/doc/html/latest/process/submitting-patches.html#sign-your-work-the-developer-s-certificate-of-origin

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH] Before adding to usablemem_rgns, check if the memory range is already included.

[chenhaixiang (A)]

> > When kexec_iomem_for_each_line() reads from /proc/iomem, concurrent
> > modifications to /proc/iomem may lead to usablemem_rgns recording
> > duplicate Crash kernel segments.
> > This can result in the number of retrieved Crash kernel segments
> > exceeding CRASH_MAX_RESERVED_RANGES, triggering a realloc of the
> > crash_reserved_mem in usablemem_rgns, which could crash the process.
> > We should ensure that each range added to usablemem_rgns is unique to
> > prevent these issues.
> 
> Thanks Chen Haixiang,
> 
> This looks reasonable to me, but, as per Kernel submissions, I will need an
> Signed-off-by line [1] in order to apply this patch.
> 
> [1]
> https://www.kernel.org/doc/html/latest/process/submitting-patches.html#sign
> -your-work-the-developer-s-certificate-of-origin

Thank you for your feedback, and I apologize for missing the Signed-off-by line. I'll resend the patch v2 with the required Signed-off-by line shortly.

Best regards,
Chen Haixiang

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec