From: Peter Zijlstra <peterz@infradead.org>
To: x86@kernel.org
Cc: linux-kernel@vger.kernel.org, Kyle Huey <me@kylehuey.com>,
Alexandre Chartre <alexandre.chartre@oracle.com>,
Robert O'Callahan <rocallahan@gmail.com>,
"Paul E. McKenney" <paulmck@kernel.org>,
Frederic Weisbecker <frederic@kernel.org>,
Paolo Bonzini <pbonzini@redhat.com>,
Sean Christopherson <sean.j.christopherson@intel.com>,
Masami Hiramatsu <mhiramat@kernel.org>,
Petr Mladek <pmladek@suse.com>,
Steven Rostedt <rostedt@goodmis.org>,
Joel Fernandes <joel@joelfernandes.org>,
Boris Ostrovsky <boris.ostrovsky@oracle.com>,
Juergen Gross <jgross@suse.com>, Brian Gerst <brgerst@gmail.com>,
Andy Lutomirski <luto@kernel.org>,
Josh Poimboeuf <jpoimboe@redhat.com>,
Peter Zijlstra <peterz@infradead.org>
Subject: [PATCH 1/7] x86/debug: Allow a single level of #DB recursion
Date: Thu, 20 Aug 2020 12:38:33 +0200 [thread overview]
Message-ID: <20200820104905.118636133@infradead.org> (raw)
In-Reply-To: 20200820103832.486877479@infradead.org
From: Andy Lutomirski <luto@kernel.org>
Trying to clear DR7 around a #DB from usermode malfunctions if we
schedule when delivering SIGTRAP. Rather than trying to define a
special no-recursion region, just allow a single level of recursion.
We do the same thing for NMI, and it hasn't caused any problems yet.
Fixes: 9f58fdde95c9 ("x86/db: Split out dr6/7 handling")
Reported-by: Kyle Huey <me@kylehuey.com>
Debugged-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/8b9bd05f187231df008d48cf818a6a311cbd5c98.1597882384.git.luto@kernel.org
---
arch/x86/kernel/traps.c | 65 ++++++++++++++++++++++--------------------------
1 file changed, 31 insertions(+), 34 deletions(-)
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -729,20 +729,9 @@ static bool is_sysenter_singlestep(struc
#endif
}
-static __always_inline void debug_enter(unsigned long *dr6, unsigned long *dr7)
+static __always_inline unsigned long debug_read_clear_dr6(void)
{
- /*
- * Disable breakpoints during exception handling; recursive exceptions
- * are exceedingly 'fun'.
- *
- * Since this function is NOKPROBE, and that also applies to
- * HW_BREAKPOINT_X, we can't hit a breakpoint before this (XXX except a
- * HW_BREAKPOINT_W on our stack)
- *
- * Entry text is excluded for HW_BP_X and cpu_entry_area, which
- * includes the entry stack is excluded for everything.
- */
- *dr7 = local_db_save();
+ unsigned long dr6;
/*
* The Intel SDM says:
@@ -755,15 +744,12 @@ static __always_inline void debug_enter(
*
* Keep it simple: clear DR6 immediately.
*/
- get_debugreg(*dr6, 6);
+ get_debugreg(dr6, 6);
set_debugreg(0, 6);
/* Filter out all the reserved bits which are preset to 1 */
- *dr6 &= ~DR6_RESERVED;
-}
+ dr6 &= ~DR6_RESERVED;
-static __always_inline void debug_exit(unsigned long dr7)
-{
- local_db_restore(dr7);
+ return dr6;
}
/*
@@ -863,6 +849,18 @@ static void handle_debug(struct pt_regs
static __always_inline void exc_debug_kernel(struct pt_regs *regs,
unsigned long dr6)
{
+ /*
+ * Disable breakpoints during exception handling; recursive exceptions
+ * are exceedingly 'fun'.
+ *
+ * Since this function is NOKPROBE, and that also applies to
+ * HW_BREAKPOINT_X, we can't hit a breakpoint before this (XXX except a
+ * HW_BREAKPOINT_W on our stack)
+ *
+ * Entry text is excluded for HW_BP_X and cpu_entry_area, which
+ * includes the entry stack is excluded for everything.
+ */
+ unsigned long dr7 = local_db_save();
bool irq_state = idtentry_enter_nmi(regs);
instrumentation_begin();
@@ -883,6 +881,8 @@ static __always_inline void exc_debug_ke
instrumentation_end();
idtentry_exit_nmi(regs, irq_state);
+
+ local_db_restore(dr7);
}
static __always_inline void exc_debug_user(struct pt_regs *regs,
@@ -894,6 +894,15 @@ static __always_inline void exc_debug_us
*/
WARN_ON_ONCE(!user_mode(regs));
+ /*
+ * NB: We can't easily clear DR7 here because
+ * idtentry_exit_to_usermode() can invoke ptrace, schedule, access
+ * user memory, etc. This means that a recursive #DB is possible. If
+ * this happens, that #DB will hit exc_debug_kernel() and clear DR7.
+ * Since we're not on the IST stack right now, everything will be
+ * fine.
+ */
+
irqentry_enter_from_user_mode(regs);
instrumentation_begin();
@@ -907,36 +916,24 @@ static __always_inline void exc_debug_us
/* IST stack entry */
DEFINE_IDTENTRY_DEBUG(exc_debug)
{
- unsigned long dr6, dr7;
-
- debug_enter(&dr6, &dr7);
- exc_debug_kernel(regs, dr6);
- debug_exit(dr7);
+ exc_debug_kernel(regs, debug_read_clear_dr6());
}
/* User entry, runs on regular task stack */
DEFINE_IDTENTRY_DEBUG_USER(exc_debug)
{
- unsigned long dr6, dr7;
-
- debug_enter(&dr6, &dr7);
- exc_debug_user(regs, dr6);
- debug_exit(dr7);
+ exc_debug_user(regs, debug_read_clear_dr6());
}
#else
/* 32 bit does not have separate entry points. */
DEFINE_IDTENTRY_RAW(exc_debug)
{
- unsigned long dr6, dr7;
-
- debug_enter(&dr6, &dr7);
+ unsigned long dr6 = debug_read_clear_dr6();
if (user_mode(regs))
exc_debug_user(regs, dr6);
else
exc_debug_kernel(regs, dr6);
-
- debug_exit(dr7);
}
#endif
next prev parent reply other threads:[~2020-08-20 10:53 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-20 10:38 [RFC][PATCH 0/7] x86/debug: Untangle handle_debug() Peter Zijlstra
2020-08-20 10:38 ` Peter Zijlstra [this message]
2020-08-26 13:54 ` [PATCH 1/7] x86/debug: Allow a single level of #DB recursion Sasha Levin
2020-08-20 10:38 ` [RFC][PATCH 2/7] x86/debug: Sync BTF earlier Peter Zijlstra
2020-08-21 2:25 ` Thomas Gleixner
2020-08-20 10:38 ` [RFC][PATCH 3/7] x86/debug: Move kprobe_debug_handler() into exc_debug_kernel() Peter Zijlstra
2020-08-21 13:54 ` Masami Hiramatsu
2020-08-20 10:38 ` [RFC][PATCH 4/7] x86/debug: Move historical SYSENTER junk " Peter Zijlstra
2020-08-20 14:45 ` Brian Gerst
2020-08-20 15:08 ` peterz
2020-08-20 15:16 ` Josh Poimboeuf
2020-08-20 15:21 ` peterz
2020-08-20 16:17 ` Josh Poimboeuf
2020-08-20 16:34 ` peterz
2020-08-20 16:43 ` Andy Lutomirski
2020-08-20 17:21 ` Josh Poimboeuf
2020-08-20 18:19 ` peterz
2020-08-20 15:28 ` Daniel Thompson
2020-08-20 15:51 ` peterz
2020-08-21 10:19 ` peterz
2020-08-21 13:31 ` Daniel Thompson
2020-08-20 10:38 ` [RFC][PATCH 5/7] x86/debug: Remove handle_debug(.user) argument Peter Zijlstra
2020-08-20 10:38 ` [RFC][PATCH 6/7] x86/debug: Simplify #DB signal code Peter Zijlstra
2020-08-20 10:38 ` [RFC][PATCH 7/7] x86/debug: Move cond_local_irq_enable() block into exc_debug_user() Peter Zijlstra
2020-08-21 2:29 ` [RFC][PATCH 0/7] x86/debug: Untangle handle_debug() Thomas Gleixner
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=20200820104905.118636133@infradead.org \
--to=peterz@infradead.org \
--cc=alexandre.chartre@oracle.com \
--cc=boris.ostrovsky@oracle.com \
--cc=brgerst@gmail.com \
--cc=frederic@kernel.org \
--cc=jgross@suse.com \
--cc=joel@joelfernandes.org \
--cc=jpoimboe@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=me@kylehuey.com \
--cc=mhiramat@kernel.org \
--cc=paulmck@kernel.org \
--cc=pbonzini@redhat.com \
--cc=pmladek@suse.com \
--cc=rocallahan@gmail.com \
--cc=rostedt@goodmis.org \
--cc=sean.j.christopherson@intel.com \
--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