public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Stefan Seyfried <seife@suse.de>
To: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: linux-kernel@vger.kernel.org, Pavel Machek <pavel@suse.cz>,
	Andrew Morton <akpm@osdl.org>
Subject: Re: 2.6.9-rc2-mm3: swsusp horribly slow on AMD64
Date: Sun, 26 Sep 2004 22:48:52 +0200	[thread overview]
Message-ID: <41572B34.3010209@suse.de> (raw)
In-Reply-To: <200409262125.38271.rjw@sisk.pl>

[-- Attachment #1: Type: text/plain, Size: 760 bytes --]

Rafael J. Wysocki wrote:

>>Try to unload all modules etc, see if it goes away.
> 
> I guess it will, but I'll check.

please try attached patch first. The comments should explain it pretty
well. It seems to have helped me: without it, sysrq-p during writing
(even if not that slow) almost always was in pccardd, now it is idling
in swapper task.
Maybe i am totally wrong but you may give it a shot.

>>If not, fix sysrq  to work for you, and look at backtrace.
> 
> This would be more time-consuming. :-)

maybe you just press wrong keys? On my Dell D600, although SysRQ is in
blue on PrtSc, no Fn-Key is needed but only ALT-PrtSc.
-- 
Stefan Seyfried, QA / R&D Team Mobile Devices, SUSE LINUX AG Nürnberg.

"Any ideas, John?"
"Well, surrounding them's out."

[-- Attachment #2: swsusp-disable-irqs-before-writing-image.diff --]
[-- Type: text/x-patch, Size: 1780 bytes --]

Only in kernel/power/: .built-in.o.cmd
Only in kernel/power/: .console.o.cmd
Only in kernel/power/: .disk.o.cmd
Only in kernel/power/: .main.o.cmd
Only in kernel/power/: .pm.o.cmd
Only in kernel/power/: .poweroff.o.cmd
Only in kernel/power/: .process.o.cmd
Only in kernel/power/: .swsusp.o.cmd
Only in kernel/power/: built-in.o
Only in kernel/power/: console.o
diff -up ../linux-2.6.8-14/kernel/power/disk.c kernel/power/disk.c
--- ../linux-2.6.8-14/kernel/power/disk.c	2004-09-23 02:22:50.000000000 +0200
+++ kernel/power/disk.c	2004-09-26 22:41:14.735864464 +0200
@@ -182,6 +182,13 @@ int pm_suspend_disk(void)
 		goto Done;
 
 	if (in_suspend) {
+		/*
+		 * during swsusp_suspend, the machine basically suspends,
+		 * does the atomic copy and then resumes again. This is
+		 * why we have to disable irqs again or there may be trouble.
+		 */
+		local_irq_disable();
+
 		pr_debug("PM: writing image.\n");
 
 		/*
Only in kernel/power/: disk.o
Only in kernel/power/: main.o
Only in kernel/power/: pm.o
Only in kernel/power/: poweroff.o
Only in kernel/power/: process.o
diff -up ../linux-2.6.8-14/kernel/power/swsusp.c kernel/power/swsusp.c
--- ../linux-2.6.8-14/kernel/power/swsusp.c	2004-09-23 02:22:51.000000000 +0200
+++ kernel/power/swsusp.c	2004-09-26 22:45:24.027966304 +0200
@@ -853,7 +853,10 @@ int swsusp_suspend(void)
 	local_irq_disable();
 	save_processor_state();
 	error = swsusp_arch_suspend();
-	/* Restore control flow magically appears here */
+	/* Restore control flow magically appears here during resume.
+	 * During suspend, this is not the end! We still have to write
+	 * the image to disk and power off.
+	 */
 	restore_processor_state();
 	restore_highmem();
 	local_irq_enable();
Only in kernel/power/: swsusp.c-orig
Only in kernel/power/: swsusp.o

  reply	other threads:[~2004-09-26 20:52 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-09-25 10:14 2.6.9-rc2-mm3: swsusp horribly slow on AMD64 Rafael J. Wysocki
2004-09-25 21:33 ` Stefan Seyfried
2004-09-26 10:02   ` Rafael J. Wysocki
2004-09-26 10:10     ` Pavel Machek
2004-09-26 11:45     ` Rafael J. Wysocki
2004-09-26 13:22       ` Pavel Machek
2004-09-26 17:06       ` Rafael J. Wysocki
2004-09-26 18:34         ` Pavel Machek
2004-09-26 19:25           ` Rafael J. Wysocki
2004-09-26 20:48             ` Stefan Seyfried [this message]
2004-09-26 22:05               ` Rafael J. Wysocki
2004-09-26 22:41               ` Pavel Machek
2004-09-25 23:40 ` Pavel Machek
2004-09-26 10:08   ` Rafael J. Wysocki
2004-09-26 10:09     ` Pavel Machek
2004-09-26 11:37       ` Rafael J. Wysocki
2004-09-26 13:20         ` Pavel Machek
2004-09-27 23:23           ` 2.6.9-rc2-mm3: swsusp horribly slow on AMD64 (update) Rafael J. Wysocki
2004-09-28  2:03             ` Narayan Desai
2004-09-29 21:58             ` 2.6.9-rc2-mm[2-4]: zaphod-scheduler.patch makes swsusp incredibly slow (was: Re: 2.6.9-rc2-mm3: swsusp horribly slow on AMD64) Rafael J. Wysocki
2004-09-29 22:50               ` Pavel Machek
2004-09-26 21:00         ` 2.6.9-rc2-mm3: swsusp horribly slow on AMD64 Stefan Seyfried
2004-09-26 22:01           ` Rafael J. Wysocki
2004-09-26 22:33             ` Pavel Machek
2004-09-27  9:11               ` Rafael J. Wysocki

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=41572B34.3010209@suse.de \
    --to=seife@suse.de \
    --cc=akpm@osdl.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pavel@suse.cz \
    --cc=rjw@sisk.pl \
    /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