linux-kernel - Re: Linux 5.5=rc1 kselftest rseq test build failure

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: <1871753014.2121.1576077544796.JavaMail.zimbra@efficios.com>
Date:   Wed, 11 Dec 2019 10:19:04 -0500 (EST)
From:   Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:     Shuah Khan <skhan@...uxfoundation.org>
Cc:     linux-kselftest <linux-kselftest@...r.kernel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: Linux 5.5=rc1 kselftest rseq test build failure

----- On Dec 10, 2019, at 7:27 PM, Shuah Khan skhan@...uxfoundation.org wrote:

> Hi Mathieu,
> 
> I am seeing rseq test build failure on Linux 5.5-rc1.
> 
> gcc -O2 -Wall -g -I./ -I../../../../usr/include/ -L./ -Wl,-rpath=./
> param_test.c -lpthread -lrseq -o ...tools/testing/selftests/rseq/param_test
> param_test.c:18:21: error: static declaration of ‘gettid’ follows
> non-static declaration
>    18 | static inline pid_t gettid(void)
>       |                     ^~~~~~
> In file included from /usr/include/unistd.h:1170,
>                  from param_test.c:11:
> /usr/include/x86_64-linux-gnu/bits/unistd_ext.h:34:16: note: previous
> declaration of ‘gettid’ was here
>    34 | extern __pid_t gettid (void) __THROW;
>       |                ^~~~~~
> make: *** [Makefile:28: ...tools/testing/selftests/rseq/param_test] Error 1
> 
> 
> The following obvious change fixes it. However, there could be reason
> why this was defined here. If you think this is the right fix, I can
> send the patch. I started seeing this with gcc version 9.2.1 20191008

This issue is caused by introduction of "gettid()" in glibc 2.30. I don't
think we want to introduce a build dependency on glibc 2.30 for kernel
selftests. Removing the gettid() symbol as you propose here will break
build environments with glibc < 2.30.

We could eventually try to figure out whether the glibc headers implement
gettid() at build time (not sure how), or we could simply rename our own
"gettid()" to "rseq_gettid()", thus removing the namespace clash with
glibc.

I can propose a patch renaming gettid() to rseq_gettid() is that approach
is OK with you.

Thanks,

Mathieu


> 
> diff --git a/tools/testing/selftests/rseq/param_test.c
> b/tools/testing/selftests/rseq/param_test.c
> index eec2663261f2..18a0fa1235a7 100644
> --- a/tools/testing/selftests/rseq/param_test.c
> +++ b/tools/testing/selftests/rseq/param_test.c
> @@ -15,11 +15,6 @@
>  #include <errno.h>
>  #include <stddef.h>
> 
> -static inline pid_t gettid(void)
> -{
> -       return syscall(__NR_gettid);
> -}
> -
> 
> thanks,
> -- Shuah

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com