Re: [PATCH -mm] coredump: format_corename: don't append .%pid if multi-threaded
From: Michael Kerrisk
Date: Tue Jul 22 2008 - 08:21:34 EST
On Tue, Jul 22, 2008 at 2:18 PM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> If the coredumping is multi-threaded, format_corename() appends .%pid
> to the corename. This was needed before the proper multi-thread core
> dump support, now all the threads in the mm go into a single unified
> core file.
>
> Remove this special case, it is not even documented and we have "%p"
> and core_uses_pid.
Hi Oleg,
I have not thought about this at any length, but one question that
jumps to mind: could this feature still be useful for LinuxThreads,
where each thread does indeed have a separate PID?
Cheers,
Michael
>
> Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
>
> --- 26-rc2/fs/exec.c~FORMAT_CORENAME_NO_MT_PID 2008-07-22 15:42:15.000000000 +0400
> +++ 26-rc2/fs/exec.c 2008-07-22 15:46:04.000000000 +0400
> @@ -1373,7 +1373,7 @@ EXPORT_SYMBOL(set_binfmt);
> * name into corename, which must have space for at least
> * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator.
> */
> -static int format_corename(char *corename, int nr_threads, long signr)
> +static int format_corename(char *corename, long signr)
> {
> const char *pat_ptr = core_pattern;
> int ispipe = (*pat_ptr == '|');
> @@ -1480,8 +1480,7 @@ static int format_corename(char *corenam
> * If core_pattern does not include a %p (as is the default)
> * and core_uses_pid is set, then .%pid will be appended to
> * the filename. Do not do this for piped commands. */
> - if (!ispipe && !pid_in_pattern
> - && (core_uses_pid || nr_threads)) {
> + if (!ispipe && !pid_in_pattern && core_uses_pid) {
> rc = snprintf(out_ptr, out_end - out_ptr,
> ".%d", task_tgid_vnr(current));
> if (rc > out_end - out_ptr)
> @@ -1745,7 +1744,7 @@ int do_coredump(long signr, int exit_cod
> * uses lock_kernel()
> */
> lock_kernel();
> - ispipe = format_corename(corename, retval, signr);
> + ispipe = format_corename(corename, signr);
> unlock_kernel();
> /*
> * Don't bother to check the RLIMIT_CORE value if core_pattern points
>
>
--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
man-pages online: http://www.kernel.org/doc/man-pages/online_pages.html
Found a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/