linux-kernel - Re: [PATCH] perf tests: Fix regex for record+probe_libc_inet

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <80d65e5c-40d7-b0a8-6f6f-c0ddb4cbb77c@linux.vnet.ibm.com>
Date:   Sat, 19 May 2018 00:58:32 +0530
From:   Sandipan Das <sandipan@...ux.vnet.ibm.com>
To:     Arnaldo Carvalho de Melo <acme@...nel.org>
Cc:     jolsa@...hat.com, linux-kernel@...r.kernel.org,
        ravi.bangoria@...ux.ibm.com, naveen.n.rao@...ux.vnet.ibm.com,
        tmricht@...ux.ibm.com
Subject: Re: [PATCH] perf tests: Fix regex for record+probe_libc_inet_pton.sh

Hi Arnaldo,

On 05/19/2018 12:51 AM, Arnaldo Carvalho de Melo wrote:
> Em Fri, May 18, 2018 at 12:54:17PM +0530, Sandipan Das escreveu:
>> This test currently fails because the regular expressions for
>> matching the output of perf script do not consider the symbol
>> offsets to be part of the output.
>>
>> The symbol offsets are seen because of the default behaviour
>> introduced by commit 4140d2ea74b3 ("perf script: Show symbol
>> offsets by default").
>>
>> Before applying this patch:
>>
>>   # perf test -v "probe libc's inet_pton & backtrace it with ping"
>>
>>   62: probe libc's inet_pton & backtrace it with ping       :
>>   --- start ---
>>   test child forked, pid 30389
>>   ping 30406 [002] 307144.280983: probe_libc:inet_pton: (7f4117adf220)
>>   7f4117adf220 __GI___inet_pton+0x0 (/usr/lib64/libc-2.25.so)
>>   FAIL: expected backtrace entry 1 ".*inet_pton[[:space:]]\(/usr/lib64/libc-2.25.so|inlined\)$" got "7f4117adf220 __GI___inet_pton+0x0 (/usr/lib64/libc-2.25.so)"
>>   test child finished with -1
>>   ---- end ----
>>   probe libc's inet_pton & backtrace it with ping: FAILED!
>>
>> After applying this patch:
>>
>>   # perf test -v "probe libc's inet_pton & backtrace it with ping"
>>
>>   62: probe libc's inet_pton & backtrace it with ping       :
>>   --- start ---
>>   test child forked, pid 30539
>>   ping 30556 [003] 307254.313217: probe_libc:inet_pton: (7fe19ab10220)
>>   7fe19ab10220 __GI___inet_pton+0x0 (/usr/lib64/libc-2.25.so)
>>   7fe19aad5ebd getaddrinfo+0x11d (/usr/lib64/libc-2.25.so)
>>   56351e3c1c71 main+0x891 (/usr/bin/ping)
>>   test child finished with 0
>>   ---- end ----
>>   probe libc's inet_pton & backtrace it with ping: Ok
> 
> Here it went from failing with:
> 
> [root@...enth ~]# perf test -v pton
> 64: probe libc's inet_pton & backtrace it with ping       :
> --- start ---
> test child forked, pid 22590
> ping 22607 [001] 12782.425689: probe_libc:inet_pton: (7f8686da4e40)
> 7f8686da4e40 __GI___inet_pton+0x0 (/usr/lib64/libc-2.26.so)
> FAIL: expected backtrace entry 1 ".*inet_pton[[:space:]]\(/usr/lib64/libc-2.26.so|inlined\)$" got "7f8686da4e40 __GI___inet_pton+0x0 (/usr/lib64/libc-2.26.so)"
> test child finished with -1
> ---- end ----
> probe libc's inet_pton & backtrace it with ping: FAILED!
> [root@...enth ~]# 
> 
> To failing with:
> 
> [root@...enth ~]# perf test -v pton
> 64: probe libc's inet_pton & backtrace it with ping       :
> --- start ---
> test child forked, pid 28954
> ping 28971 [002] 14277.711200: probe_libc:inet_pton: (7fc9d66e3e40)
> 7fc9d66e3e40 __GI___inet_pton+0x0 (/usr/lib64/libc-2.26.so)
> 7fc9d66b02b4 getaddrinfo+0x124 (/usr/lib64/libc-2.26.so)
> 56075cb98f40 [unknown] (/usr/bin/ping)
> FAIL: expected backtrace entry 3 ".*\+0x[[:xdigit:]]+[[:space:]]\(.*/bin/ping.*\)$" got "56075cb98f40 [unknown] (/usr/bin/ping)"
> test child finished with -1
> ---- end ----
> probe libc's inet_pton & backtrace it with ping: FAILED!
> [root@...enth ~]# 
> 
> Trying to figure this out...
> 

Looks like perf failed to resolve the symbol's name for the last entry
in the callchain. I did not consider this case. So, if this happens, we
would be better off using the original regex for the last line:

expected[3]=".*\(.*/bin/ping.*\)$"

- Sandipan