lkml.org 
[lkml]   [2012]   [Jun]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[PATCH 2/3 v2] x86: Remove cmpxchg from i386 NMI nesting code
From: Steven Rostedt <srostedt@redhat.com>

I've been informed by someone on LWN called 'slashdot' that
some i386 machines do not support a true cmpxchg. The cmpxchg
used by the i386 NMI nesting code must be a true cmpxchg as
disabling interrupts will not work for NMIs (which is the work
around for i386s that do not have a true cmpxchg).

This 'slashdot' character also suggested a fix to the issue.
As the state of the nesting NMIs goes as follows:

NOT_RUNNING -> EXECUTING
EXECUTING -> NOT_RUNNING
EXECUTING -> LATCHED
LATCHED -> EXECUTING

Having these states as enum values of:

NOT_RUNNING = 0
EXECUTING = 1
LATCHED = 2

Instead of a cmpxchg to make EXECUTING -> NOT_RUNNING a
dec_and_test() would work as well. If the dec_and_test brings
the state to NOT_RUNNING, that is the same as a cmpxchg
succeeding to change EXECUTING to NOT_RUNNING. If a nested NMI
were to come in and change it to LATCHED, the dec_and_test() would
convert the state to EXECUTING (what we want it to be in such a
case anyway).

I asked 'slashdot' to post this as a patch, but it never came to
be. I decided to do the work instead.

Thanks to H. Peter Anvin for suggesting to use this_cpu_dec_and_return()
instead of local_dec_and_test(&__get_cpu_var()).

Link: http://lwn.net/Articles/484932/

Cc: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
arch/x86/kernel/nmi.c | 35 +++++++++++++++++++++--------------
1 file changed, 21 insertions(+), 14 deletions(-)
diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c
index a0b2f84..a15a888 100644
--- a/arch/x86/kernel/nmi.c
+++ b/arch/x86/kernel/nmi.c
@@ -365,8 +365,9 @@ static __kprobes void default_do_nmi(struct pt_regs *regs)
#ifdef CONFIG_X86_32
/*
* For i386, NMIs use the same stack as the kernel, and we can
- * add a workaround to the iret problem in C. Simply have 3 states
- * the NMI can be in.
+ * add a workaround to the iret problem in C (preventing nested
+ * NMIs if an NMI takes a trap). Simply have 3 states the NMI
+ * can be in:
*
* 1) not running
* 2) executing
@@ -383,13 +384,20 @@ static __kprobes void default_do_nmi(struct pt_regs *regs)
* If an NMI hits a breakpoint that executes an iret, another
* NMI can preempt it. We do not want to allow this new NMI
* to run, but we want to execute it when the first one finishes.
- * We set the state to "latched", and the first NMI will perform
- * an cmpxchg on the state, and if it doesn't successfully
- * reset the state to "not running" it will restart the next
- * NMI.
+ * We set the state to "latched", and the exit of the first NMI will
+ * perform a dec_return, if the result is zero (NOT_RUNNING), then
+ * it will simply exit the NMI handler. If not, the dec_return
+ * would have set the state to NMI_EXECUTING (what we want it to
+ * be when we are running). In this case, we simply jump back
+ * to rerun the NMI handler again, and restart the 'latched' NMI.
+ *
+ * No trap (breakpoint or page fault) should be hit before nmi_restart,
+ * thus there is no race between the first check of state for NOT_RUNNING
+ * and setting it to NMI_EXECUTING. The HW will prevent nested NMIs
+ * at this point.
*/
enum nmi_states {
- NMI_NOT_RUNNING,
+ NMI_NOT_RUNNING = 0,
NMI_EXECUTING,
NMI_LATCHED,
};
@@ -397,18 +405,17 @@ static DEFINE_PER_CPU(enum nmi_states, nmi_state);

#define nmi_nesting_preprocess(regs) \
do { \
- if (__get_cpu_var(nmi_state) != NMI_NOT_RUNNING) { \
- __get_cpu_var(nmi_state) = NMI_LATCHED; \
+ if (this_cpu_read(nmi_state) != NMI_NOT_RUNNING) { \
+ this_cpu_write(nmi_state, NMI_LATCHED); \
return; \
} \
- nmi_restart: \
- __get_cpu_var(nmi_state) = NMI_EXECUTING; \
- } while (0)
+ this_cpu_write(nmi_state, NMI_EXECUTING); \
+ } while (0); \
+ nmi_restart:

#define nmi_nesting_postprocess() \
do { \
- if (cmpxchg(&__get_cpu_var(nmi_state), \
- NMI_EXECUTING, NMI_NOT_RUNNING) != NMI_EXECUTING) \
+ if (this_cpu_dec_return(nmi_state)) \
goto nmi_restart; \
} while (0)
#else /* x86_64 */
--
1.7.10

[unhandled content-type:application/pgp-signature]
\
 
 \ /
  Last update: 2012-06-09 04:41    [from the cache]
©2003-2014 Jasper Spaans. Advertise on this site