| 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
| ||
|
Date: Fri, 02 Dec 2011 10:46:03 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: John Kacur <jkacur@...hat.com>
Cc: Thomas Gleixner <tglx@...utronix.de>,
rt-users <linux-rt-users@...r.kernel.org>,
lkml <linux-kernel@...r.kernel.org>,
Andi Kleen <andi@...stfloor.org>
Subject: Re: [PATCH-RT] Fix for sleeping function called from invalid
context when running gdb
On Fri, 2011-12-02 at 14:39 +0100, John Kacur wrote:
> While running gdb on cyclictest on v3.2-rc2-rt3 I got the following error
>
> [root@...rship ~]# [11887.443716] BUG: sleeping function called from invalid con
> text at /home/jkacur/linux-rt/kernel/rtmutex.c:645
> [11887.443720] in_atomic(): 1, irqs_disabled(): 0, pid: 4859, name: cyclictest
> [11887.443723] no locks held by cyclictest/4859.
> [11887.443726] Pid: 4859, comm: cyclictest Not tainted 3.2.0-rc2-rt3-debug-expr+
> #1
> [11887.443728] Call Trace:
> [11887.443737] [<ffffffff81035543>] __might_sleep+0x185/0x18d
> [11887.443743] [<ffffffff8107c3ed>] rt_spin_lock_fastlock.clone.0+0x24/0x31
> [11887.443749] [<ffffffff813fba9b>] rt_spin_lock+0x16/0x40
> [11887.443754] [<ffffffff81053eee>] force_sig_info+0x3d/0xeb
> [11887.443758] [<ffffffff81053fb2>] force_sig+0x16/0x18
> [11887.443761] [<ffffffff813fd0a1>] do_trap+0xef/0x130
> [11887.443766] [<ffffffff81076525>] ? trace_hardirqs_on_caller+0x12a/0x161
> [11887.443771] [<ffffffff813fd16f>] do_int3+0x8d/0x9a
> [11887.443774] [<ffffffff813fc9e7>] int3+0x27/0x40
>
> I found that there is a patch for this in v2.6.33.9-rt31 that I forward-ported
> to v3.2-rc2-rt3
>
> I believe that Thomas Gleixner is the original author, but I'm not sure.
> Thomas - can you provide your signed-off-by if you are the author?
Actually, the author was Andi. It's in the 2.6.24-rt patch set too.
Here's the change log:
>From ak@...e.de Sat Oct 27 10:32:13 2007
Date: Sat, 27 Oct 2007 12:39:33 +0200
From: Andi Kleen <ak@...e.de>
To: linux-rt-users@...r.kernel.org
Subject: [PATCH] Don't disable preemption in exception handlers without IST
Some of the exception handlers that run on an IST in a normal kernel
still disable preemption. This causes might_sleep warning when sending signals
for debugging in PREEMPT-RT because sending signals can take a lock.
Since the ISTs are disabled now for those don't disable the preemption.
This completes the remove IST patch I sent some time ago and fixes
another case where using gdb caused warnings.
Also it will likely improve latency a little bit.
Signed-off-by: Andi Kleen <ak@...e.de>
Thomas, I think this patch is worth adding, unless you have a better
work around.
-- Steve
>
> Forward-ported from v2.6.33.9-rt31 to v3.2-rc2-rt3
> Signed-off-by: John Kacur <jkacur@...hat.com>
> ---
> arch/x86/kernel/traps.c | 24 +++++++++++++-----------
> 1 files changed, 13 insertions(+), 11 deletions(-)
>
> diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
> index a8e3eb8..3654390 100644
> --- a/arch/x86/kernel/traps.c
> +++ b/arch/x86/kernel/traps.c
> @@ -87,9 +87,10 @@ static inline void conditional_sti(struct pt_regs *regs)
> local_irq_enable();
> }
>
> -static inline void preempt_conditional_sti(struct pt_regs *regs)
> +static inline void preempt_conditional_sti(struct pt_regs *regs, int stack)
> {
> - inc_preempt_count();
> + if (stack)
> + inc_preempt_count();
> if (regs->flags & X86_EFLAGS_IF)
> local_irq_enable();
> }
> @@ -100,11 +101,12 @@ static inline void conditional_cli(struct pt_regs *regs)
> local_irq_disable();
> }
>
> -static inline void preempt_conditional_cli(struct pt_regs *regs)
> +static inline void preempt_conditional_cli(struct pt_regs *regs, int stack)
> {
> if (regs->flags & X86_EFLAGS_IF)
> local_irq_disable();
> - dec_preempt_count();
> + if (stack)
> + dec_preempt_count();
> }
>
> static void __kprobes
> @@ -222,9 +224,9 @@ dotraplinkage void do_stack_segment(struct pt_regs *regs, long error_code)
> if (notify_die(DIE_TRAP, "stack segment", regs, error_code,
> 12, SIGBUS) == NOTIFY_STOP)
> return;
> - preempt_conditional_sti(regs);
> + preempt_conditional_sti(regs, STACKFAULT_STACK);
> do_trap(12, SIGBUS, "stack segment", regs, error_code, NULL);
> - preempt_conditional_cli(regs);
> + preempt_conditional_cli(regs, STACKFAULT_STACK);
> }
>
> dotraplinkage void do_double_fault(struct pt_regs *regs, long error_code)
> @@ -316,9 +318,9 @@ dotraplinkage void __kprobes do_int3(struct pt_regs *regs, long error_code)
> return;
> #endif
>
> - preempt_conditional_sti(regs);
> + preempt_conditional_sti(regs, DEBUG_STACK);
> do_trap(3, SIGTRAP, "int3", regs, error_code, NULL);
> - preempt_conditional_cli(regs);
> + preempt_conditional_cli(regs, DEBUG_STACK);
> }
>
> #ifdef CONFIG_X86_64
> @@ -412,12 +414,12 @@ dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code)
> return;
>
> /* It's safe to allow irq's after DR6 has been saved */
> - preempt_conditional_sti(regs);
> + preempt_conditional_sti(regs, DEBUG_STACK);
>
> if (regs->flags & X86_VM_MASK) {
> handle_vm86_trap((struct kernel_vm86_regs *) regs,
> error_code, 1);
> - preempt_conditional_cli(regs);
> + preempt_conditional_cli(regs, DEBUG_STACK);
> return;
> }
>
> @@ -436,7 +438,7 @@ dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code)
> si_code = get_si_code(tsk->thread.debugreg6);
> if (tsk->thread.debugreg6 & (DR_STEP | DR_TRAP_BITS) || user_icebp)
> send_sigtrap(tsk, regs, error_code, si_code);
> - preempt_conditional_cli(regs);
> + preempt_conditional_cli(regs, DEBUG_STACK);
>
> return;
> }
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists