From: Fengguang Wu <fengguang.wu@intel.com>
To: Jiri Kosina <jkosina@suse.cz>
Cc: "H. Peter Anvin" <hpa@zytor.com>,
"H. Peter Anvin" <hpa@linux.intel.com>,
linux-kernel@vger.kernel.org
Subject: Re: [x86] Kernel panic - not syncing: Fatal exception in interrupt
Date: Mon, 22 Jul 2013 09:07:56 +0800 [thread overview]
Message-ID: <20130722010756.GA3365@localhost> (raw)
In-Reply-To: <alpine.LNX.2.00.1307211004100.31632@pobox.suse.cz>
Hi Jiri,
> What I am however wondering whether can't be case here is that the jump
> label was used before int3_notifier has been registered.
> I am thinking about ways around this, but we'll probably have to do the
> same ftrace is doing, i.e. hook into do_int3() directly instead of relying
> on the notifier to be registered in time.
>
> Fengguang, as I am not able to reproduce this bug locally, could you do me
> a favor and test whether the patch below works the problem around, just
> for the sake of testing the hypothesis?
I tested 1000 boots with the patch and find no more boot problem.
Thanks,
Fengguang
> From: Jiri Kosina <jkosina@suse.cz>
> Subject: [PATCH] x86: call out into int3 handler directly instead of using notifier
>
> ---
> arch/x86/include/asm/alternative.h | 2 ++
> arch/x86/kernel/alternative.c | 22 +++++++++++++++++++++-
> arch/x86/kernel/traps.c | 4 ++++
> 3 files changed, 27 insertions(+), 1 deletions(-)
>
> diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h
> index 3abf8dd..c22a41d 100644
> --- a/arch/x86/include/asm/alternative.h
> +++ b/arch/x86/include/asm/alternative.h
> @@ -5,6 +5,7 @@
> #include <linux/stddef.h>
> #include <linux/stringify.h>
> #include <asm/asm.h>
> +#include <asm/ptrace.h>
>
> /*
> * Alternative inline assembly for SMP.
> @@ -232,6 +233,7 @@ struct text_poke_param {
> size_t len;
> };
>
> +extern int poke_bp_int3_handler(struct pt_regs *regs);
> extern void *text_poke(void *addr, const void *opcode, size_t len);
> extern void *text_poke_bp(void *addr, const void *opcode, size_t len, void *handler);
> extern void *text_poke_smp(void *addr, const void *opcode, size_t len);
> diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
> index 0ab4936..e1088f2 100644
> --- a/arch/x86/kernel/alternative.c
> +++ b/arch/x86/kernel/alternative.c
> @@ -605,6 +605,24 @@ static void do_sync_core(void *info)
> static bool bp_patching_in_progress;
> static void *bp_int3_handler, *bp_int3_addr;
>
> +int poke_bp_int3_handler(struct pt_regs *regs)
> +{
> + /* bp_patching_in_progress */
> + smp_rmb();
> +
> + if (likely(!bp_patching_in_progress))
> + return 0;
> +
> + if (user_mode_vm(regs) || regs->ip != (unsigned long)bp_int3_addr)
> + return 0;
> +
> + /* set up the specified breakpoint handler */
> + regs->ip = (unsigned long) bp_int3_handler;
> +
> + return 1;
> +
> +}
> +
> static int int3_notify(struct notifier_block *self, unsigned long val, void *data)
> {
> struct die_args *args = data;
> @@ -689,6 +707,7 @@ void *text_poke_bp(void *addr, const void *opcode, size_t len, void *handler)
> return addr;
> }
>
> +#if 0
> /* this one needs to run before anything else handles it as a
> * regular exception */
> static struct notifier_block int3_nb = {
> @@ -700,8 +719,9 @@ static int __init int3_init(void)
> {
> return register_die_notifier(&int3_nb);
> }
> -
> arch_initcall(int3_init);
> +#endif
> +
> /*
> * Cross-modifying kernel text with stop_machine().
> * This code originally comes from immediate value.
> diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
> index 772e2a8..e464764 100644
> --- a/arch/x86/kernel/traps.c
> +++ b/arch/x86/kernel/traps.c
> @@ -58,6 +58,7 @@
> #include <asm/mce.h>
> #include <asm/fixmap.h>
> #include <asm/mach_traps.h>
> +#include <asm/alternative.h>
>
> #ifdef CONFIG_X86_64
> #include <asm/x86_init.h>
> @@ -324,6 +325,9 @@ dotraplinkage void __kprobes notrace do_int3(struct pt_regs *regs, long error_co
> ftrace_int3_handler(regs))
> return;
> #endif
> + if (poke_bp_int3_handler(regs))
> + return;
> +
> prev_state = exception_enter();
> #ifdef CONFIG_KGDB_LOW_LEVEL_TRAP
> if (kgdb_ll_trap(DIE_INT3, "int3", regs, error_code, X86_TRAP_BP,
>
> --
> Jiri Kosina
> SUSE Labs
next prev parent reply other threads:[~2013-07-22 1:08 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-20 13:12 [x86] Kernel panic - not syncing: Fatal exception in interrupt Fengguang Wu
2013-07-21 0:23 ` H. Peter Anvin
2013-07-21 8:30 ` Jiri Kosina
2013-07-21 16:21 ` Fengguang Wu
2013-07-21 17:21 ` H. Peter Anvin
2013-07-22 11:18 ` Jiri Kosina
2013-07-22 1:07 ` Fengguang Wu [this message]
2013-07-22 11:14 ` [PATCH -tip/x86/jumplabel] x86: call out into int3 handler directly instead of using notifier Jiri Kosina
2013-07-22 11:24 ` Masami Hiramatsu
2013-07-22 20:53 ` H. Peter Anvin
2013-07-22 21:00 ` Jiri Kosina
2013-07-23 1:24 ` Masami Hiramatsu
2013-07-23 7:45 ` Ingo Molnar
2013-07-23 8:09 ` Jiri Kosina
2013-07-24 3:55 ` [tip:perf/core] kprobes/x86: Call out into INT3 " tip-bot for Jiri Kosina
2013-07-24 14:33 ` H. Peter Anvin
2013-07-29 9:06 ` Jiri Kosina
2013-07-29 9:14 ` H. Peter Anvin
2013-07-29 9:18 ` Jiri Kosina
2013-07-29 9:23 ` H. Peter Anvin
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=20130722010756.GA3365@localhost \
--to=fengguang.wu@intel.com \
--cc=hpa@linux.intel.com \
--cc=hpa@zytor.com \
--cc=jkosina@suse.cz \
--cc=linux-kernel@vger.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).