From: Denys Vlasenko <dvlasenk@redhat.com>
To: Andy Lutomirski <luto@kernel.org>, Takashi Iwai <tiwai@suse.de>,
Stefan Seyfried <stefan.seyfried@googlemail.com>,
X86 ML <x86@kernel.org>
Cc: Jiri Kosina <jkosina@suse.cz>,
LKML <linux-kernel@vger.kernel.org>, Tejun Heo <tj@kernel.org>
Subject: Re: [PATCH] x86, entry: Check for syscall exit work with IRQs disabled
Date: Mon, 23 Mar 2015 20:21:33 +0100 [thread overview]
Message-ID: <551067BD.7060006@redhat.com> (raw)
In-Reply-To: <189320d42b4d671df78c10555976bb10af1ffc75.1427137498.git.luto@kernel.org>
On 03/23/2015 08:10 PM, Andy Lutomirski wrote:
> We currently have a race: if we're preempted during syscall exit, we
> can fail to process syscall return work that is queued up while
> we're preempted in ret_from_sys_call after checking ti.flags.
>
> Fix it by disabling interrupts before checking ti.flags.
>
> Fixes: 96b6352c1271 x86_64, entry: Remove the syscall exit audit and schedule optimizations
> Reported-by: Stefan Seyfried <stefan.seyfried@googlemail.com>
> Reported-by: Takashi Iwai <tiwai@suse.de>
> Signed-off-by: Andy Lutomirski <luto@kernel.org>
> ---
>
> Ingo, I don't understand the LOCKDEP_SYS_EXIT stuff. Can you take a quick
> look to confirm that it's okay to call it more than once?
>
> arch/x86/kernel/entry_64.S | 18 ++++++++++++++----
> 1 file changed, 14 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
> index 1d74d161687c..2babb393915e 100644
> --- a/arch/x86/kernel/entry_64.S
> +++ b/arch/x86/kernel/entry_64.S
> @@ -364,12 +364,21 @@ system_call_fastpath:
> * Has incomplete stack frame and undefined top of stack.
> */
> ret_from_sys_call:
> - testl $_TIF_ALLWORK_MASK,TI_flags+THREAD_INFO(%rsp,RIP-ARGOFFSET)
> - jnz int_ret_from_sys_call_fixup /* Go the the slow path */
> -
> LOCKDEP_SYS_EXIT
> DISABLE_INTERRUPTS(CLBR_NONE)
> TRACE_IRQS_OFF
> +
> + /*
> + * We must check ti flags with interrupts (or at least preemption)
> + * off because we must *never* return to userspace without
> + * processing exit work that is enqueued if we're preempted here.
> + * In particular, returning to userspace with any of the one-shot
> + * flags (TIF_NOTIFY_RESUME, TIF_USER_RETURN_NOTIFY, etc) set is
> + * very bad.
> + */
> + testl $_TIF_ALLWORK_MASK,TI_flags+THREAD_INFO(%rsp,RIP-ARGOFFSET)
> + jnz int_ret_from_sys_call_fixup /* Go the the slow path */
^^^^^^^^^^^^^^^^^^^^
typo here; s/the the/to the/
> +
> CFI_REMEMBER_STATE
> /*
> * sysretq will re-enable interrupts:
> @@ -386,7 +395,7 @@ ret_from_sys_call:
>
> int_ret_from_sys_call_fixup:
> FIXUP_TOP_OF_STACK %r11, -ARGOFFSET
> - jmp int_ret_from_sys_call
> + jmp int_ret_from_sys_call_irqs_off
>
> /* Do syscall tracing */
> tracesys:
> @@ -432,6 +441,7 @@ tracesys_phase2:
> GLOBAL(int_ret_from_sys_call)
> DISABLE_INTERRUPTS(CLBR_NONE)
> TRACE_IRQS_OFF
> +int_ret_from_sys_call_irqs_off:
> movl $_TIF_ALLWORK_MASK,%edi
> /* edi: mask to check */
> GLOBAL(int_with_check)
You can avoid having to know LOCKDEP_SYS_EXIT :)
Just set %edi = $_TIF_ALLWORK_MASK, and jump a bit farther:
movl $_TIF_ALLWORK_MASK,%edi
testl %edi,TI_flags+THREAD_INFO(%rsp,RIP-ARGOFFSET)
jnz int_ret_from_sys_call_fixup /* Go to the slow path */
...
...
GLOBAL(int_ret_from_sys_call)
DISABLE_INTERRUPTS(CLBR_NONE)
TRACE_IRQS_OFF
movl $_TIF_ALLWORK_MASK,%edi
/* edi: mask to check */
GLOBAL(int_with_check)
LOCKDEP_SYS_EXIT_IRQ
int_ret_from_sys_call_irqs_off: <========== HERE
next prev parent reply other threads:[~2015-03-23 19:21 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-15 8:17 PANIC: double fault, error_code: 0x0 in 4.0.0-rc3-2, kvm related? Stefan Seyfried
2015-03-18 14:16 ` Takashi Iwai
2015-03-18 15:05 ` Takashi Iwai
2015-03-18 17:43 ` Takashi Iwai
2015-03-18 17:46 ` Takashi Iwai
2015-03-18 18:03 ` Andy Lutomirski
2015-03-18 19:03 ` Stefan Seyfried
2015-03-18 19:26 ` Andy Lutomirski
2015-03-18 20:05 ` Stefan Seyfried
2015-03-18 20:51 ` Andy Lutomirski
2015-03-18 21:12 ` Stefan Seyfried
2015-03-18 21:21 ` Andy Lutomirski
2015-03-18 21:41 ` Stefan Seyfried
2015-03-18 21:49 ` Denys Vlasenko
2015-03-18 21:53 ` Stefan Seyfried
2015-03-18 20:06 ` Denys Vlasenko
2015-03-18 20:49 ` Andy Lutomirski
2015-03-18 21:06 ` Denys Vlasenko
2015-03-18 21:17 ` Andy Lutomirski
2015-03-18 21:32 ` Linus Torvalds
2015-03-18 21:42 ` Denys Vlasenko
2015-03-18 21:55 ` Andy Lutomirski
2015-03-18 22:17 ` Denys Vlasenko
2015-03-18 22:20 ` Andy Lutomirski
2015-03-18 22:27 ` Denys Vlasenko
2015-03-18 22:18 ` Linus Torvalds
2015-03-18 22:24 ` Andy Lutomirski
2015-03-18 22:22 ` Jiri Kosina
2015-03-18 22:28 ` Linus Torvalds
2015-03-18 22:29 ` Andy Lutomirski
2015-03-18 22:29 ` Andy Lutomirski
2015-03-18 22:38 ` Stefan Seyfried
2015-03-18 22:40 ` Andy Lutomirski
2015-03-18 23:22 ` Andy Lutomirski
2015-03-19 0:23 ` Stefan Seyfried
2015-03-19 0:57 ` Andy Lutomirski
2015-03-19 2:15 ` Linus Torvalds
2015-03-19 6:24 ` Stefan Seyfried
2015-03-19 10:16 ` Takashi Iwai
2015-03-19 10:58 ` Denys Vlasenko
2015-03-19 11:21 ` Takashi Iwai
2015-03-19 12:48 ` Denys Vlasenko
2015-03-19 13:47 ` Takashi Iwai
2015-03-19 14:55 ` Takashi Iwai
2015-03-19 15:22 ` Takashi Iwai
2015-03-19 15:41 ` Andy Lutomirski
2015-03-19 15:51 ` Takashi Iwai
2015-03-19 16:01 ` Andy Lutomirski
2015-03-20 18:16 ` Denys Vlasenko
2015-03-20 18:50 ` Takashi Iwai
2015-03-23 9:02 ` Takashi Iwai
2015-03-23 9:35 ` Takashi Iwai
2015-03-23 13:22 ` Takashi Iwai
2015-03-23 16:07 ` Denys Vlasenko
2015-03-23 17:18 ` Takashi Iwai
2015-03-23 17:46 ` Denys Vlasenko
2015-03-23 18:43 ` Takashi Iwai
2015-03-23 18:38 ` Andy Lutomirski
2015-03-23 18:48 ` Andy Lutomirski
2015-03-23 18:59 ` Takashi Iwai
2015-03-23 19:10 ` [PATCH] x86, entry: Check for syscall exit work with IRQs disabled Andy Lutomirski
2015-03-23 19:21 ` Denys Vlasenko [this message]
2015-03-23 19:27 ` Andy Lutomirski
2015-03-23 19:32 ` Andy Lutomirski
2015-03-24 11:17 ` Takashi Iwai
2015-03-24 20:08 ` Ingo Molnar
2015-03-25 0:35 ` Andy Lutomirski
2015-03-25 12:21 ` Ingo Molnar
2015-03-25 15:07 ` Andy Lutomirski
2015-03-25 9:13 ` [tip:x86/asm] x86/asm/entry: " tip-bot for Andy Lutomirski
2015-03-23 18:54 ` PANIC: double fault, error_code: 0x0 in 4.0.0-rc3-2, kvm related? Stefan Seyfried
2015-03-23 18:56 ` Takashi Iwai
2015-03-23 19:07 ` Denys Vlasenko
2015-03-23 19:10 ` Andy Lutomirski
2015-03-19 13:21 ` Denys Vlasenko
2015-03-18 21:49 ` Stefan Seyfried
2015-03-28 23:57 ` Maciej W. Rozycki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=551067BD.7060006@redhat.com \
--to=dvlasenk@redhat.com \
--cc=jkosina@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=stefan.seyfried@googlemail.com \
--cc=tiwai@suse.de \
--cc=tj@kernel.org \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).