From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Luming Yu" Date: Thu, 22 May 2008 09:16:28 +0000 Subject: Re: [RFC PATCH] set TASK_TRACED before arch_ptrace code to fix a race Message-Id: <3877989d0805220216o5add20ddye2a1fde98a0c1e69@mail.gmail.com> List-Id: References: <3877989d0805211947i54bacc7cv619541e9b40824fb@mail.gmail.com> <1211446045.5610.33.camel@elijah.suse.cz> In-Reply-To: <1211446045.5610.33.camel@elijah.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Petr Tesarik Cc: LKML , linux-ia64@vger.kernel.org, Roland McGrath On Thu, May 22, 2008 at 4:47 PM, Petr Tesarik wrote: > On Thu, 2008-05-22 at 10:47 +0800, Luming Yu wrote: >> Hello list, >> >> The following patch is to fixed a race in ptrace_stop handling which >> causes "strace" hang if the target process blocks SIGTRAP with the >> test case filed at >> https://bugzilla.redhat.com/show_bug.cgi?idD6200#c16. >> Please note this is just IA64 problem because just IA64 has >> arch_ptrace_stop_needed defined, and has arch_ptrace_stop defined that >> would set notify_resume flags for syncing rbs...but it also opens the >> door to invoke ia64_do_signal->get_signal_to_deliver before setting >> current PTRACED flag. Please help review. >> >> **The patch is enclosed in text attachment* >> **I'm using web client to send the patch* * > > I'm inlining the patch for sake of convenience: > thanks. >> >> Signed-off-by: Yu Luming >> -------------------------------------- >> signal.c | 5 +++-- >> 1 file changed, 3 insertions(+), 2 deletions(-) >> >> --- 0/kernel/signal.c 2008-05-14 02:24:51.000000000 +0800 >> +++ 1/kernel/signal.c 2008-05-22 13:54:42.000000000 +0800 >> @@ -1488,6 +1488,9 @@ >> { >> int killed = 0; >> >> + /* Let the debugger run. */ >> + __set_current_state(TASK_TRACED); >> + > > That's probably not what we want. What happens if the task then sleeps > during the user-space access? Unless I forgot something obvious, it will > never get scheduled again... My intention is to disable signal delivering before TASK_TRACED flag is set for correctly handling ptrace_stop() with SIGTRAP masked. Although this patch totally is a hack, but it should clearly shows where the problem is that I want to solve.. > > Petr Tesarik > >> if (arch_ptrace_stop_needed(exit_code, info)) { >> /* >> * The arch code has something special to do before a >> @@ -1516,8 +1519,6 @@ >> current->last_siginfo = info; >> current->exit_code = exit_code; >> >> - /* Let the debugger run. */ >> - __set_current_state(TASK_TRACED); >> spin_unlock_irq(¤t->sighand->siglock); >> read_lock(&tasklist_lock); >> if (!unlikely(killed) && may_ptrace_stop()) { > >