From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: Maxim Levitsky <maximlevitsky@gmail.com>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: linux-pm@lists.linux-foundation.org,
LKML <linux-kernel@vger.kernel.org>,
linux-mm <linux-mm@kvack.org>,
Andrew Morton <akpm@linux-foundation.org>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>
Subject: Re: [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] Memory allocations in .suspend became very unreliable)
Date: Sat, 30 Jan 2010 19:47:10 +0100 [thread overview]
Message-ID: <201001301947.10453.rjw@sisk.pl> (raw)
In-Reply-To: <1264866419.27933.0.camel@maxim-laptop>
On Saturday 30 January 2010, Maxim Levitsky wrote:
> On Sun, 2010-01-17 at 01:38 +0100, Rafael J. Wysocki wrote:
> > Hi,
> >
> > I thing the snippet below is a good summary of what this is about.
>
> Any progress on that?
Well, I'm waiting for you to report back:
http://patchwork.kernel.org/patch/74740/
The patch is appended once again for convenience.
Rafael
---
From: Rafael J. Wysocki <rjw@sisk.pl>
Subject: MM / PM: Force GFP_NOIO during suspend/hibernation and resume
There are quite a few GFP_KERNEL memory allocations made during
suspend/hibernation and resume that may cause the system to hang,
because the I/O operations they depend on cannot be completed due to
the underlying devices being suspended.
Avoid this problem by clearing the __GFP_IO and __GFP_FS bits in
gfp_allowed_mask before suspend/hibernation and restoring the
original values of these bits in gfp_allowed_mask durig the
subsequent resume.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Reported-by: Maxim Levitsky <maximlevitsky@gmail.com>
---
kernel/power/hibernate.c | 6 ++++++
kernel/power/power.h | 3 +++
kernel/power/suspend.c | 2 ++
mm/Makefile | 1 +
mm/pm.c | 28 ++++++++++++++++++++++++++++
5 files changed, 40 insertions(+)
Index: linux-2.6/kernel/power/hibernate.c
===================================================================
--- linux-2.6.orig/kernel/power/hibernate.c
+++ linux-2.6/kernel/power/hibernate.c
@@ -334,6 +334,7 @@ int hibernation_snapshot(int platform_mo
goto Close;
suspend_console();
+ mm_force_noio_allocations();
error = dpm_suspend_start(PMSG_FREEZE);
if (error)
goto Recover_platform;
@@ -351,6 +352,7 @@ int hibernation_snapshot(int platform_mo
dpm_resume_end(in_suspend ?
(error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE);
+ mm_allow_io_allocations();
resume_console();
Close:
platform_end(platform_mode);
@@ -448,11 +450,13 @@ int hibernation_restore(int platform_mod
pm_prepare_console();
suspend_console();
+ mm_force_noio_allocations();
error = dpm_suspend_start(PMSG_QUIESCE);
if (!error) {
error = resume_target_kernel(platform_mode);
dpm_resume_end(PMSG_RECOVER);
}
+ mm_allow_io_allocations();
resume_console();
pm_restore_console();
return error;
@@ -481,6 +485,7 @@ int hibernation_platform_enter(void)
entering_platform_hibernation = true;
suspend_console();
+ mm_force_noio_allocations();
error = dpm_suspend_start(PMSG_HIBERNATE);
if (error) {
if (hibernation_ops->recover)
@@ -518,6 +523,7 @@ int hibernation_platform_enter(void)
Resume_devices:
entering_platform_hibernation = false;
dpm_resume_end(PMSG_RESTORE);
+ mm_allow_io_allocations();
resume_console();
Close:
Index: linux-2.6/kernel/power/power.h
===================================================================
--- linux-2.6.orig/kernel/power/power.h
+++ linux-2.6/kernel/power/power.h
@@ -187,6 +187,9 @@ static inline void suspend_test_finish(c
#ifdef CONFIG_PM_SLEEP
/* kernel/power/main.c */
extern int pm_notifier_call_chain(unsigned long val);
+/* mm/pm.c */
+extern void mm_force_noio_allocations(void);
+extern void mm_allow_io_allocations(void);
#endif
#ifdef CONFIG_HIGHMEM
Index: linux-2.6/kernel/power/suspend.c
===================================================================
--- linux-2.6.orig/kernel/power/suspend.c
+++ linux-2.6/kernel/power/suspend.c
@@ -208,6 +208,7 @@ int suspend_devices_and_enter(suspend_st
goto Close;
}
suspend_console();
+ mm_force_noio_allocations();
suspend_test_start();
error = dpm_suspend_start(PMSG_SUSPEND);
if (error) {
@@ -224,6 +225,7 @@ int suspend_devices_and_enter(suspend_st
suspend_test_start();
dpm_resume_end(PMSG_RESUME);
suspend_test_finish("resume devices");
+ mm_allow_io_allocations();
resume_console();
Close:
if (suspend_ops->end)
Index: linux-2.6/mm/Makefile
===================================================================
--- linux-2.6.orig/mm/Makefile
+++ linux-2.6/mm/Makefile
@@ -40,3 +40,4 @@ obj-$(CONFIG_MEMORY_FAILURE) += memory-f
obj-$(CONFIG_HWPOISON_INJECT) += hwpoison-inject.o
obj-$(CONFIG_DEBUG_KMEMLEAK) += kmemleak.o
obj-$(CONFIG_DEBUG_KMEMLEAK_TEST) += kmemleak-test.o
+obj-$(CONFIG_PM_SLEEP) += pm.o
Index: linux-2.6/mm/pm.c
===================================================================
--- /dev/null
+++ linux-2.6/mm/pm.c
@@ -0,0 +1,28 @@
+#include <linux/gfp.h>
+
+static gfp_t saved_gfp_allowed_mask;
+
+#define GFP_IOFS (__GFP_IO | __GFP_FS)
+
+/**
+ * mm_force_noio_allocations - Modify gfp_allowed_mask to disable IO allocations
+ *
+ * Change gfp_allowed_mask by unsetting __GFP_IO and __GFP_FS in it and save the
+ * old value.
+ */
+void mm_force_noio_allocations(void)
+{
+ saved_gfp_allowed_mask = gfp_allowed_mask;
+ gfp_allowed_mask &= ~GFP_IOFS;
+}
+
+/**
+ * mm_allow_io_allocations - Modify gfp_allowed_mask to allow IO allocations
+ *
+ * If the saved value of gfp_allowed_mask has __GFP_IO set, modify the current
+ * gfp_allowed_mask by setting this bit and anlogously for __GFP_FS.
+ */
+void mm_allow_io_allocations(void)
+{
+ gfp_allowed_mask |= saved_gfp_allowed_mask & GFP_IOFS;
+}
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2010-01-30 18:46 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <1263549544.3112.10.camel@maxim-laptop>
[not found] ` <1263678289.4276.4.camel@maxim-laptop>
[not found] ` <201001162317.39940.rjw@sisk.pl>
2010-01-17 0:38 ` [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] Memory allocations in .suspend became very unreliable) Rafael J. Wysocki
2010-01-17 1:24 ` Oliver Neukum
2010-01-17 13:27 ` Rafael J. Wysocki
2010-01-17 13:36 ` Rafael J. Wysocki
2010-01-17 18:58 ` Benjamin Herrenschmidt
2010-01-17 23:00 ` Rafael J. Wysocki
2010-01-18 7:53 ` Oliver Neukum
2010-01-18 16:17 ` [linux-pm] [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: " Alan Stern
2010-01-18 20:59 ` Rafael J. Wysocki
2010-01-19 15:13 ` Alan Stern
2010-01-18 20:56 ` [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] " Rafael J. Wysocki
2010-01-18 21:55 ` Benjamin Herrenschmidt
2010-01-18 23:33 ` Rafael J. Wysocki
2010-01-17 13:55 ` Rafael J. Wysocki
2010-01-17 16:21 ` [linux-pm] [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: " Minchan Kim
2010-01-17 16:23 ` Minchan Kim
2010-01-18 0:25 ` Rafael J. Wysocki
2010-01-18 2:20 ` KOSAKI Motohiro
2010-01-18 21:06 ` Rafael J. Wysocki
2010-01-19 9:15 ` Oliver Neukum
2010-01-19 20:34 ` Rafael J. Wysocki
2010-01-18 17:00 ` [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] " Oliver Neukum
2010-01-18 20:41 ` Rafael J. Wysocki
2010-01-19 9:25 ` Oliver Neukum
2010-01-19 20:37 ` Rafael J. Wysocki
2010-01-20 14:05 ` Oliver Neukum
2010-01-20 21:13 ` Rafael J. Wysocki
2010-01-18 2:16 ` KOSAKI Motohiro
2010-01-18 20:55 ` Rafael J. Wysocki
2010-01-19 1:19 ` KOSAKI Motohiro
2010-01-19 3:19 ` Benjamin Herrenschmidt
2010-01-19 9:04 ` Bastien ROUCARIES
2010-01-19 23:17 ` Benjamin Herrenschmidt
2010-01-20 11:31 ` Oliver Neukum
2010-01-20 21:11 ` Benjamin Herrenschmidt
2010-01-20 21:12 ` Rafael J. Wysocki
2010-01-19 20:47 ` Rafael J. Wysocki
2010-01-20 0:33 ` KOSAKI Motohiro
2010-01-20 21:21 ` Rafael J. Wysocki
2010-01-21 0:47 ` KOSAKI Motohiro
2010-01-21 20:21 ` Rafael J. Wysocki
2010-01-21 20:42 ` [linux-pm] [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: " Nigel Cunningham
2010-01-21 21:38 ` Rafael J. Wysocki
2010-01-22 1:31 ` [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] " KOSAKI Motohiro
2010-01-22 1:42 ` KOSAKI Motohiro
2010-01-22 10:11 ` Maxim Levitsky
2010-01-22 21:19 ` [Update][PATCH] MM / PM: Force GFP_NOIO during suspend/hibernation and resume Rafael J. Wysocki
2010-01-23 9:29 ` Maxim Levitsky
2010-01-25 21:49 ` Rafael J. Wysocki
2010-01-25 21:52 ` Alexey Starikovskiy
2010-01-30 18:56 ` Rafael J. Wysocki
2010-01-30 20:42 ` Maxim Levitsky
2010-01-30 20:53 ` Rafael J. Wysocki
2010-01-22 20:58 ` [RFC][PATCH] PM: Force GFP_NOIO during suspend/resume (was: Re: [linux-pm] Memory allocations in .suspend became very unreliable) Rafael J. Wysocki
2010-01-30 15:46 ` Maxim Levitsky
2010-01-30 18:47 ` Rafael J. Wysocki [this message]
2010-01-30 20:37 ` Maxim Levitsky
2010-02-01 19:51 ` Maxim Levitsky
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=201001301947.10453.rjw@sisk.pl \
--to=rjw@sisk.pl \
--cc=akpm@linux-foundation.org \
--cc=benh@kernel.crashing.org \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-pm@lists.linux-foundation.org \
--cc=maximlevitsky@gmail.com \
/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).