Linux RAID subsystem development
 help / color / mirror / Atom feed
* [PATCH] Fix NULL difference in super_by_fd
@ 2022-12-12  9:23 lixiaokeng
  2022-12-13 15:30 ` Jes Sorensen
  0 siblings, 1 reply; 4+ messages in thread
From: lixiaokeng @ 2022-12-12  9:23 UTC (permalink / raw)
  To: jes, linux-raid; +Cc: linfeilong, liuzhiqiang (I), Wu Guanghao

When we create 100 partitions(major is 259 not 254) in a raid device,
mdadm may coredump:

Core was generated by `/usr/sbin/mdadm --detail --export /dev/md1p3'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  __strlen_sse2 ()
    at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
126		movdqu	(%rax), %xmm4
(gdb) bt
#0  __strlen_sse2 ()
    at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
#1  0x00007f1944659139 in __strcpy_chk (
    dest=dest@entry=0x55ea8d7c23ac "", src=0x0,
    destlen=destlen@entry=32) at strcpy_chk.c:28
#2  0x000055ea8d10b66d in strcpy (__src=<optimized out>,
    __dest=0x55ea8d7c23ac "")
    at /usr/include/bits/string_fortified.h:79
#3  super_by_fd (fd=fd@entry=3,
    subarrayp=subarrayp@entry=0x7ffe6a1dff08) at util.c:1289
#4  0x000055ea8d11b3a6 in Detail (
    dev=0x7ffe6a1e2f22 "/dev/md1p3", c=0x7ffe6a1e1700)
    at Detail.c:101
#5  0x000055ea8d101e61 in misc_list (c=<optimized out>,
    ss=<optimized out>, dump_directory=<optimized out>,
    ident=<optimized out>, devlist=<optimized out>)
    at mdadm.c:1959
#6  main (argc=<optimized out>, argv=<optimized out>)
    at mdadm.c:1629

The direct cause is fd2devnm return NULL. Here add a check.

Signed-off-by:Lixiaokeng<lixiaokeng@huawei.com>
Signed-off-by:Wuguanghao<wuguanghao3@huawei.com>
---
 util.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/util.c b/util.c
index 26ffdcea..843bfc6d 100644
--- a/util.c
+++ b/util.c
@@ -1177,6 +1177,7 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
 	int i;
 	char *subarray = NULL;
 	char container[32] = "";
+	char *devnm = NULL;

 	sra = sysfs_read(fd, NULL, GET_VERSION);

@@ -1222,7 +1223,10 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
 		if (subarrayp)
 			*subarrayp = subarray;
 		strcpy(st->container_devnm, container);
-		strcpy(st->devnm, fd2devnm(fd));
+		if (devnm = fd2devnm(fd))
+			strcpy(st->devnm, devnm);
+		else
+			st->devnm[0] = '\0';
 	} else
 		free(subarray);

-- 

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix NULL difference in super_by_fd
  2022-12-12  9:23 [PATCH] Fix NULL difference in super_by_fd lixiaokeng
@ 2022-12-13 15:30 ` Jes Sorensen
  2022-12-14  1:19   ` lixiaokeng
  2022-12-14  3:08   ` lixiaokeng
  0 siblings, 2 replies; 4+ messages in thread
From: Jes Sorensen @ 2022-12-13 15:30 UTC (permalink / raw)
  To: lixiaokeng, linux-raid; +Cc: linfeilong, liuzhiqiang (I), Wu Guanghao

On 12/12/22 04:23, lixiaokeng wrote:
> When we create 100 partitions(major is 259 not 254) in a raid device,
> mdadm may coredump:
> 
> Core was generated by `/usr/sbin/mdadm --detail --export /dev/md1p3'.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0  __strlen_sse2 ()
>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
> 126		movdqu	(%rax), %xmm4
> (gdb) bt
> #0  __strlen_sse2 ()
>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
> #1  0x00007f1944659139 in __strcpy_chk (
>     dest=dest@entry=0x55ea8d7c23ac "", src=0x0,
>     destlen=destlen@entry=32) at strcpy_chk.c:28
> #2  0x000055ea8d10b66d in strcpy (__src=<optimized out>,
>     __dest=0x55ea8d7c23ac "")
>     at /usr/include/bits/string_fortified.h:79
> #3  super_by_fd (fd=fd@entry=3,
>     subarrayp=subarrayp@entry=0x7ffe6a1dff08) at util.c:1289
> #4  0x000055ea8d11b3a6 in Detail (
>     dev=0x7ffe6a1e2f22 "/dev/md1p3", c=0x7ffe6a1e1700)
>     at Detail.c:101
> #5  0x000055ea8d101e61 in misc_list (c=<optimized out>,
>     ss=<optimized out>, dump_directory=<optimized out>,
>     ident=<optimized out>, devlist=<optimized out>)
>     at mdadm.c:1959
> #6  main (argc=<optimized out>, argv=<optimized out>)
>     at mdadm.c:1629
> 
> The direct cause is fd2devnm return NULL. Here add a check.
> 
> Signed-off-by:Lixiaokeng<lixiaokeng@huawei.com>
> Signed-off-by:Wuguanghao<wuguanghao3@huawei.com>
> ---
>  util.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/util.c b/util.c
> index 26ffdcea..843bfc6d 100644
> --- a/util.c
> +++ b/util.c
> @@ -1177,6 +1177,7 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>  	int i;
>  	char *subarray = NULL;
>  	char container[32] = "";
> +	char *devnm = NULL;
> 
>  	sra = sysfs_read(fd, NULL, GET_VERSION);
> 
> @@ -1222,7 +1223,10 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>  		if (subarrayp)
>  			*subarrayp = subarray;
>  		strcpy(st->container_devnm, container);
> -		strcpy(st->devnm, fd2devnm(fd));
> +		if (devnm = fd2devnm(fd))
> +			strcpy(st->devnm, devnm);
> +		else
> +			st->devnm[0] = '\0';

I don't think this is the correct fix. You end up returning an
incomplete 'st' entry, which could cause unexpected behavior. I think
the right way to handle this is to fail properly and return NULL from
super_by_fd(), after cleaning up properly.

Cheers,
Jes



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix NULL difference in super_by_fd
  2022-12-13 15:30 ` Jes Sorensen
@ 2022-12-14  1:19   ` lixiaokeng
  2022-12-14  3:08   ` lixiaokeng
  1 sibling, 0 replies; 4+ messages in thread
From: lixiaokeng @ 2022-12-14  1:19 UTC (permalink / raw)
  To: Jes Sorensen, linux-raid; +Cc: linfeilong, liuzhiqiang (I), Wu Guanghao



On 2022/12/13 23:30, Jes Sorensen wrote:
> On 12/12/22 04:23, lixiaokeng wrote:
>> When we create 100 partitions(major is 259 not 254) in a raid device,
>> mdadm may coredump:
>>
>> Core was generated by `/usr/sbin/mdadm --detail --export /dev/md1p3'.
>> Program terminated with signal SIGSEGV, Segmentation fault.
>> #0  __strlen_sse2 ()
>>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
>> 126		movdqu	(%rax), %xmm4
>> (gdb) bt
>> #0  __strlen_sse2 ()
>>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
>> #1  0x00007f1944659139 in __strcpy_chk (
>>     dest=dest@entry=0x55ea8d7c23ac "", src=0x0,
>>     destlen=destlen@entry=32) at strcpy_chk.c:28
>> #2  0x000055ea8d10b66d in strcpy (__src=<optimized out>,
>>     __dest=0x55ea8d7c23ac "")
>>     at /usr/include/bits/string_fortified.h:79
>> #3  super_by_fd (fd=fd@entry=3,
>>     subarrayp=subarrayp@entry=0x7ffe6a1dff08) at util.c:1289
>> #4  0x000055ea8d11b3a6 in Detail (
>>     dev=0x7ffe6a1e2f22 "/dev/md1p3", c=0x7ffe6a1e1700)
>>     at Detail.c:101
>> #5  0x000055ea8d101e61 in misc_list (c=<optimized out>,
>>     ss=<optimized out>, dump_directory=<optimized out>,
>>     ident=<optimized out>, devlist=<optimized out>)
>>     at mdadm.c:1959
>> #6  main (argc=<optimized out>, argv=<optimized out>)
>>     at mdadm.c:1629
>>
>> The direct cause is fd2devnm return NULL. Here add a check.
>>
>> Signed-off-by:Lixiaokeng<lixiaokeng@huawei.com>
>> Signed-off-by:Wuguanghao<wuguanghao3@huawei.com>
>> ---
>>  util.c | 6 +++++-
>>  1 file changed, 5 insertions(+), 1 deletion(-)
>>
>> diff --git a/util.c b/util.c
>> index 26ffdcea..843bfc6d 100644
>> --- a/util.c
>> +++ b/util.c
>> @@ -1177,6 +1177,7 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>>  	int i;
>>  	char *subarray = NULL;
>>  	char container[32] = "";
>> +	char *devnm = NULL;
>>
>>  	sra = sysfs_read(fd, NULL, GET_VERSION);
>>
>> @@ -1222,7 +1223,10 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>>  		if (subarrayp)
>>  			*subarrayp = subarray;
>>  		strcpy(st->container_devnm, container);
>> -		strcpy(st->devnm, fd2devnm(fd));
>> +		if (devnm = fd2devnm(fd))
>> +			strcpy(st->devnm, devnm);
>> +		else
>> +			st->devnm[0] = '\0';
> 
> I don't think this is the correct fix. You end up returning an
> incomplete 'st' entry, which could cause unexpected behavior. I think
> the right way to handle this is to fail properly and return NULL from
> super_by_fd(), after cleaning up properly.
> 
> Cheers,
> Jes
> 
Thanks for reply, I will change my patch and send it again.

> 
> .
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix NULL difference in super_by_fd
  2022-12-13 15:30 ` Jes Sorensen
  2022-12-14  1:19   ` lixiaokeng
@ 2022-12-14  3:08   ` lixiaokeng
  1 sibling, 0 replies; 4+ messages in thread
From: lixiaokeng @ 2022-12-14  3:08 UTC (permalink / raw)
  To: Jes Sorensen, linux-raid; +Cc: linfeilong, liuzhiqiang (I), Wu Guanghao



On 2022/12/13 23:30, Jes Sorensen wrote:
> On 12/12/22 04:23, lixiaokeng wrote:
>> When we create 100 partitions(major is 259 not 254) in a raid device,
>> mdadm may coredump:
>>
>> Core was generated by `/usr/sbin/mdadm --detail --export /dev/md1p3'.
>> Program terminated with signal SIGSEGV, Segmentation fault.
>> #0  __strlen_sse2 ()
>>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
>> 126		movdqu	(%rax), %xmm4
>> (gdb) bt
>> #0  __strlen_sse2 ()
>>     at ../sysdeps/x86_64/multiarch/strlen-vec.S:126
>> #1  0x00007f1944659139 in __strcpy_chk (
>>     dest=dest@entry=0x55ea8d7c23ac "", src=0x0,
>>     destlen=destlen@entry=32) at strcpy_chk.c:28
>> #2  0x000055ea8d10b66d in strcpy (__src=<optimized out>,
>>     __dest=0x55ea8d7c23ac "")
>>     at /usr/include/bits/string_fortified.h:79
>> #3  super_by_fd (fd=fd@entry=3,
>>     subarrayp=subarrayp@entry=0x7ffe6a1dff08) at util.c:1289
>> #4  0x000055ea8d11b3a6 in Detail (
>>     dev=0x7ffe6a1e2f22 "/dev/md1p3", c=0x7ffe6a1e1700)
>>     at Detail.c:101
>> #5  0x000055ea8d101e61 in misc_list (c=<optimized out>,
>>     ss=<optimized out>, dump_directory=<optimized out>,
>>     ident=<optimized out>, devlist=<optimized out>)
>>     at mdadm.c:1959
>> #6  main (argc=<optimized out>, argv=<optimized out>)
>>     at mdadm.c:1629
>>
>> The direct cause is fd2devnm return NULL. Here add a check.
>>
>> Signed-off-by:Lixiaokeng<lixiaokeng@huawei.com>
>> Signed-off-by:Wuguanghao<wuguanghao3@huawei.com>
>> ---
>>  util.c | 6 +++++-
>>  1 file changed, 5 insertions(+), 1 deletion(-)
>>
>> diff --git a/util.c b/util.c
>> index 26ffdcea..843bfc6d 100644
>> --- a/util.c
>> +++ b/util.c
>> @@ -1177,6 +1177,7 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>>  	int i;
>>  	char *subarray = NULL;
>>  	char container[32] = "";
>> +	char *devnm = NULL;
>>
>>  	sra = sysfs_read(fd, NULL, GET_VERSION);
>>
>> @@ -1222,7 +1223,10 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
>>  		if (subarrayp)
>>  			*subarrayp = subarray;
>>  		strcpy(st->container_devnm, container);
>> -		strcpy(st->devnm, fd2devnm(fd));
>> +		if (devnm = fd2devnm(fd))
>> +			strcpy(st->devnm, devnm);
>> +		else
>> +			st->devnm[0] = '\0';
> 
> I don't think this is the correct fix. You end up returning an
> incomplete 'st' entry, which could cause unexpected behavior. I think
> the right way to handle this is to fail properly and return NULL from
> super_by_fd(), after cleaning up properly.

Here is another question: Should devid2devnm() be modifed?
Becasue the major of MD partition may not be 254(mdp).

> 
> Cheers,
> Jes
> 
> 
> .
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2022-12-14  3:08 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-12-12  9:23 [PATCH] Fix NULL difference in super_by_fd lixiaokeng
2022-12-13 15:30 ` Jes Sorensen
2022-12-14  1:19   ` lixiaokeng
2022-12-14  3:08   ` lixiaokeng

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox