Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello Rainer, On Sat, Mar 07, 2015 at 06:47:31PM +0100, Rainer Dorsch wrote: Hi Emil, Am Samstag 07 März 2015, 14:26:17 schrieb Emil Goode: Just to clarify, suspend/resume is working but it takes more than 60 sec for suspend to complete? yes, suspend/resume is working. I did three cycles and measured the time for suspend: The first suspend took approx. 60 seconds, the second and third approx. 30 seconds. I see, it looks like its an issue with the ATA subsystem that is stalling the suspend process. [ 181.992086] ata3.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen [ 181.992090] ata3.00: failed command: FLUSH CACHE [ 181.992098] ata3.00: cmd e7/00:00:00:00:00/00:00:00:00:00/a0 tag 16 res 40/00:c8:30:0e:40/84:00:09:00:00/40 Emask 0x4 (timeout) [ 181.992100] ata3.00: status: { DRDY } [ 181.992106] ata3: hard resetting link [ 181.992108] ata3: nv: skipping hardreset on occupied port [ 182.460049] ata3: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 182.476330] ata3.00: configured for UDMA/133 [ 182.476333] ata3.00: retrying FLUSH 0xe7 Emask 0x4 [ 182.476963] ata3: EH complete [ 182.477033] ata3.00: Enabling discard_zeroes_data [ 182.477598] sd 2:0:0:0: [sda] Stopping disk [ 182.736087] PM: suspend of devices complete after 61889.003 msecs Since it is a regression what you could do is use git bisect to try and locate the commit that introduced the issue. But its time consuming and it doesn't always work. Other than that I think you should post a bug in the kernel bugzilla. Best regards, Emil -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150308000522.GA11955@x201
Bug#774036: Re: Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello Rainer, On Thu, Mar 05, 2015 at 10:09:04PM +0100, Rainer Dorsch wrote: Hi Emil, sorry for the long silence, but I did not have physical access to the machine for which I reported the bug and I did not want to install new kernel from remote... I recompiled a recent kernel with PM_TRACE_RTC enabled, but still get no log data in pm-suspend.log: root@Silberkiste:~# uname -a Linux Silberkiste 4.0.0-rc2+ #1 SMP Thu Mar 5 01:01:25 CET 2015 x86_64 GNU/Linux So the issue has not been resolved in the 4.0-rc2 kernel? root@Silberkiste:~# grep PM_TRACE_RTC /boot/config-4.0.0-rc2+ CONFIG_PM_TRACE_RTC=y root@Silberkiste:~# head -4 /usr/lib/pm-utils/pm-functions #!/bin/sh # vim:noexpandtab export PM_DEBUG=true root@Silberkiste:~# cat /var/log/pm-suspend.log root@Silberkiste:~# ls -l /var/log/pm-suspend.log -rw-r--r-- 1 root root 0 Jan 3 15:52 /var/log/pm-suspend.log root@Silberkiste:~# I did 501 echo 1 /sys/power/pm_trace 502 echo mem /sys/power/state but this did not create any logs. Do you have any suggestions, what I might do wrong? Sorry, the /var/log/pm-suspend.log is used by the pm-suspend script. Try suspending with pm-suspend and post the log. Then run the following and post the dmesg output after resuming. echo 1 /sys/power/pm_trace echo mem /sys/power/state Also did you try to suspend after blacklisting the wifi modules? I can take a look at it but note that it may require more skilled upstream people to resolve this issue. Best regards, Emil Goode Thanks, Rainer On Sunday 04 January 2015 18:12:44 Emil Goode wrote: Hello Rainer, On Sat, Jan 03, 2015 at 04:23:44PM +0100, Rainer Dorsch wrote: Hi Emil, I have only remote access to the system for the next few weeks and I feel somewhat uncomfortable to experiment with my self-compiled kernels on the system or changing networking configurations. So please apologize that I may need 3-4 weeks to test that. I assume you suggest to follow http://kernel-handbook.alioth.debian.org/ch-common-tasks.html to enable PM_TRACE_RTC and rebuild the kernel, correct? Yes and you can clone this tree for testing upstream kernels. git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git Best regards, Emil Goode On Tuesday 30 December 2014 12:40:53 Emil Goode wrote: Hello Rainer, On Tue, Dec 30, 2014 at 12:15:31AM +0100, Rainer Dorsch wrote: Hi Emil, Am Montag 29 Dezember 2014, 14:02:55 schrieb Emil Goode: Hello Rainer, The below procedure would give more information: Add the following line to the top of the file /usr/lib/pm-utils/pm-functions. export PM_DEBUG=true Then suspend with: echo 1 /sys/power/pm_trace echo mem /sys/power/state HmmI do not have a pm_trace Ok, that is because PM_TRACE_RTC needs to be enabled in the kernel config. Please be aware that when using pm_trace it will mess with your machine's real-time clock. There is a similar bug report already for an older kernel, but that seems to be fixed (?) https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=638264 The iwlagn module is for intel wireless and I believe you are using broadcom. Post dmesg and the contents of /var/log/pm-suspend.log. Have you tried disabling the mac80211 module? Can you tell how I unload a module for suspend or if http://superuser.com/questions/620201/how-to-force-ubuntu-debian-mint- to-u nload-modules-at-suspend-to-disk also applies to systemd? I think you can temporarily blacklist modules as described here: https://wiki.debian.org/KernelModuleBlacklisting I would also try the latest upstream kernel, this might have been fixed. Best regards, Emil Goode -- Rainer Dorsch http://bokomoko.de/ -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150306183637.GA2116@lianli
Bug#774036: Re: Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello Rainer, On Sat, Jan 03, 2015 at 04:23:44PM +0100, Rainer Dorsch wrote: Hi Emil, I have only remote access to the system for the next few weeks and I feel somewhat uncomfortable to experiment with my self-compiled kernels on the system or changing networking configurations. So please apologize that I may need 3-4 weeks to test that. I assume you suggest to follow http://kernel-handbook.alioth.debian.org/ch-common-tasks.html to enable PM_TRACE_RTC and rebuild the kernel, correct? Yes and you can clone this tree for testing upstream kernels. git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git Best regards, Emil Goode On Tuesday 30 December 2014 12:40:53 Emil Goode wrote: Hello Rainer, On Tue, Dec 30, 2014 at 12:15:31AM +0100, Rainer Dorsch wrote: Hi Emil, Am Montag 29 Dezember 2014, 14:02:55 schrieb Emil Goode: Hello Rainer, The below procedure would give more information: Add the following line to the top of the file /usr/lib/pm-utils/pm-functions. export PM_DEBUG=true Then suspend with: echo 1 /sys/power/pm_trace echo mem /sys/power/state HmmI do not have a pm_trace Ok, that is because PM_TRACE_RTC needs to be enabled in the kernel config. Please be aware that when using pm_trace it will mess with your machine's real-time clock. There is a similar bug report already for an older kernel, but that seems to be fixed (?) https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=638264 The iwlagn module is for intel wireless and I believe you are using broadcom. Post dmesg and the contents of /var/log/pm-suspend.log. Have you tried disabling the mac80211 module? Can you tell how I unload a module for suspend or if http://superuser.com/questions/620201/how-to-force-ubuntu-debian-mint-to-u nload-modules-at-suspend-to-disk also applies to systemd? I think you can temporarily blacklist modules as described here: https://wiki.debian.org/KernelModuleBlacklisting I would also try the latest upstream kernel, this might have been fixed. Best regards, Emil Goode -- Rainer Dorsch http://bokomoko.de/ -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150104171244.GA344@lianli
Bug#774036: Re: Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello Rainer, On Tue, Dec 30, 2014 at 12:15:31AM +0100, Rainer Dorsch wrote: Hi Emil, Am Montag 29 Dezember 2014, 14:02:55 schrieb Emil Goode: Hello Rainer, The below procedure would give more information: Add the following line to the top of the file /usr/lib/pm-utils/pm-functions. export PM_DEBUG=true Then suspend with: echo 1 /sys/power/pm_trace echo mem /sys/power/state HmmI do not have a pm_trace Ok, that is because PM_TRACE_RTC needs to be enabled in the kernel config. Please be aware that when using pm_trace it will mess with your machine's real-time clock. There is a similar bug report already for an older kernel, but that seems to be fixed (?) https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=638264 The iwlagn module is for intel wireless and I believe you are using broadcom. Post dmesg and the contents of /var/log/pm-suspend.log. Have you tried disabling the mac80211 module? Can you tell how I unload a module for suspend or if http://superuser.com/questions/620201/how-to-force-ubuntu-debian-mint-to-unload-modules-at-suspend-to-disk also applies to systemd? I think you can temporarily blacklist modules as described here: https://wiki.debian.org/KernelModuleBlacklisting I would also try the latest upstream kernel, this might have been fixed. Best regards, Emil Goode -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141230114053.GA2030@lianli
Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello Rainer, The below procedure would give more information: Add the following line to the top of the file /usr/lib/pm-utils/pm-functions. export PM_DEBUG=true Then suspend with: echo 1 /sys/power/pm_trace echo mem /sys/power/state Post dmesg and the contents of /var/log/pm-suspend.log. Have you tried disabling the mac80211 module? Best regards, Emil Goode -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141229130255.GA3410@lianli
Bug#774036: linux-image-3.16.0-4-amd64: Untainted Kernel shows similar issue
Hello, On Sun, Dec 28, 2014 at 11:15:21AM +0100, Rainer Dorsch wrote: Package: src:linux Version: 3.16.7-ckt2-1 Followup-For: Bug #774036 Dear Ben, many thanks for the superfast response and thanks for pointing out that I was running the nvidia driver. I run now with nouveau and also purge the virtualbox stuff. Unfortunately, the behavior is exactly the same though (except that there is no blinking cursor with nouveau, when the system hangs), i.e. the resume works when I use freezer in debug mode, the system hangs when I go to the next level (devices). When you get a black screen does the system hang completely or is it possible to login to the system over ssh? If it's possible to login it could be useful to post the dmesg output here. Best regards, Emil Goode -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141228145628.GA2215@lianli
Bug#746602: skb_copy_ubufs() not exported by the Debian Linux kernel 3.2.57-3
Hello Jarno, Yes it seems that the following commit should be included in 3.2. commit dcc0fb782b3a6e2abfeaaeb45dd88ed09596be0f Author: Michael S. Tsirkin m...@redhat.com Date: Fri Jul 20 09:23:20 2012 + skbuff: export skb_copy_ubufs It applies cleanly on top of 3.2.58. However, the skb_orphan_frags inline function is not in include/linux/skbuff.h of kernel 3.2.58, but it is available in the debian linux-headers-3.2.0-4-amd64 package. It seems like the following commit has been backported to the debian 3.2 kernel but not to mainline 3.2.x. commit a353e0ce0fd42d8859260666d1e9b10f2abd4698 Author: Michael S. Tsirkin m...@redhat.com Date: Fri Jul 20 09:23:07 2012 + skbuff: add an api to orphan frags Probably these need to be included together. Best regards, Emil Goode On Thu, May 01, 2014 at 01:45:28PM -0700, Jarno Rajahalme wrote: Package: linux-source-3.2 Version: 3.2.57-3 The symbol skb_copy_ubufs() is not exported in net/core/skbuff.c, even though it is exposed in include/linux/skbuff.h and used by skb_orphan_frags(). OVS kernel module fails to start due to this: Error: Module openvswitch is not currently loaded ERROR: could not insert 'openvswitch': Unknown symbol in module, or unknown parameter (see dmesg) dmesg output: [ 3315.819202] openvswitch: Unknown symbol skb_copy_ubufs (err 0) /lib/modules/3.2.0-4-amd64/source/include/linux/skbuff.h has this: static inline int skb_orphan_frags(struct sk_buff *skb, gfp_t gfp_mask) { if (likely(!(skb_shinfo(skb)-tx_flags SKBTX_DEV_ZEROCOPY))) return 0; return skb_copy_ubufs(skb, gfp_mask); } But /usr/src/linux-source-3.2/net/core/skbuff.c only has this: int skb_copy_ubufs(struct sk_buff *skb, gfp_t gfp_mask) { int i; [snipped] return 0; } Note the missing EXPORT_SYMBOL() after the function definition. It seems that skb_copy_ubufs() has been exported in Linux 3.6, as the skb_orphan_frags() was introduced. Now Debian 3.2.57-3 has a backported skb_orphan_frags(), but the export of the symbol sib_copy_ubufs() was not backported. Regards, Jarno Rajahalme -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/c7b639a4-ae25-4247-a105-d7a164964...@nicira.com -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140503000154.GA9262@lianli
Bug#721871: i915: add quirk to invert backlight brightness
Hello, I got a black screen with wheezy 3.2.55 kernel on a HP G62 Notebook, but the issue is resolved in kernel 3.12.9. Could you test the attached patch with kernel 3.2.55 that adds the invert brightness quirk? It seems that adding the quirk is the way to deal with this issue with older kernel versions. Best regards, Emil Goode From 77ab7c26d3b886869b034b8752eca092f12e90c9 Mon Sep 17 00:00:00 2001 From: Emil Goode emilgo...@gmail.com Date: Tue, 18 Feb 2014 18:09:50 +0100 Subject: [PATCH] drm/i915: add quirk to invert backlight brightness The HP G62 Notebook and Lenovo IdeaPad Yoga 11S boot to a black screen and need the invert brightness quirk. Signed-off-by: Emil Goode emilgo...@gmail.com --- drivers/gpu/drm/i915/intel_display.c |6 ++ 1 file changed, 6 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 61b708b..54507c5 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -8932,6 +8932,12 @@ struct intel_quirk intel_quirks[] = { /* Acer/Packard Bell NCL20 */ { 0x2a42, 0x1025, 0x034b, quirk_invert_brightness }, + /* HP G62 Notebook */ + { 0x0046, 0x103c, 0x1439, quirk_invert_brightness }, + + /* Lenovo IdeaPad Yoga 11S */ + { 0x0166, 0x17aa, 0x3977, quirk_invert_brightness }, + /* Dell XPS13 HD Sandy Bridge */ { 0x0116, 0x1028, 0x052e, quirk_no_pcm_pwm_enable }, /* Dell XPS13 HD and XPS13 FHD Ivy Bridge */ -- 1.7.10.4