gmane.linux.power-management.general

[RFC] Tying sysfs "wakeup" file to platform-level handler

Deepak Saxena — 2008-11-24T20:01:22

Hi, On the OLPC XO laptop, we want a way to enable/disable wakeup events for a device from userspace at suspend time depending on the method we are using for suspending. For example, if the system auto-suspends, we want to wake up on input from the PS2 device; if the user shuts the lid, we want to wakeup on lid state change events; if the user presses the power button to suspend, we only want to wakeup on another power button event. We have currently exported a set of files via sysfs in /sys/power/wakeup_events as such: bash-3.2# ls /sys/power/wakeup_events/ ac_power battery_error battery_state ps2event all battery_soc ebook_mode_change wlan Writing a 1 to one of these files will enable that device as a wake up source in the Embedded Controller (EC). Writing a 0 will disable it. See [1] for code. This interface works but I would like to replace it with something that can be pushed upstream. Each of the devices we care about already has an entry in the sysfs tree or could hav

2.6.28-rc6-git1: Reported regressions from 2.6.27

Rafael J. Wysocki — 2008-11-22T20:24:49

This message contains a list of some regressions from 2.6.27, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know. If you know of any other unresolved regressions from 2.6.27, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid. Each entry from the list will be sent additionally in an automatic reply to this message with CCs to the people involved in reporting and handling the issue. Listed regressions statistics: Date Total Pending Unresolved ---------------------------------------- 2008-11-22 93 25 15 2008-11-16 89 32 18 2008-11-09 73 40 27 2008-11-02 55 41 29 2008-10-25 26 25 20 Unresolved regressions ---------------------- Bug-Entry: http://bugzilla.kernel.org/show_bug.cgi?id=12064 Subject: [regression] Measured 688 cycles TSC warp;

Crash during suspend to disk

Simon Holm Thøgersen — 2008-11-20T12:41:56

Hi I hope this information is sufficient to debug my problem, though it unfortunately is somewhat incomplete. I took two pictures of the OOPS with my mobile phone, but unfortunately the picture of the upper part is so blurred that it is unreadable. Apparently I have to be in a hurry getting out of the door for this to happen. Here is what I got from the lower part Call Trace: __blkdev_put swsusp_write hibernate state_store state_store state_store kobj_attr_store sysfs_write_file sysfs_write_file vfs_write sys_write sysenter_do_call Code: ...... EIP: [<...>] iput ---[ end trace ........................... ]--- There is a lot of blanks here that I should be able provide on request as well. The problem occured in my third attempt in quick succession to suspend to disk. The first two did not succeed due to insufficient free swap. I got 768 MB ram and a ~1 GB swap partition. I believe top said something like Mem: 3xxxxxk used and Swap: 6xxxxxk used before the third attempt to suspend t

New comer in the linux-pm list

David GIGUET — 2008-11-20T12:40:56

Info on hibernation swap space requirement

Francesco VIRLINZI — 2008-11-18T08:59:30

Hi all. I added the hibernation support on a sh4 based platform and it works fine, but I don't understand how big the swap space has to be. My system has 64Mb and I was assuming, I can freeze ~32 Mb therefore I was creating a swap space of 40/50 Mb. But it seems it isn't enough, the system is requiring more space. Could you say me how evaluate the right (minimal) swap space required when I know the DRam memory size. Regards Francesco

[PATCH] ACPI suspend: Blacklist boxes that require us toset SCI_EN directly on resume

Rafael J. Wysocki — 2008-11-17T22:33:12

From: Rafael J. Wysocki sisk.pl> Subject: ACPI suspend: Blacklist boxes that require us to set SCI_EN directly on resume Some Apple boxes evidently require us to set SCI_EN on resume directly, because if we don't do that, they hung somewhere in the resume code path. Moreover, on these boxes it is not sufficient to use acpi_enable() to turn ACPI on during resume. All of this is against the ACPI specification which states that (1) the BIOS is supposed to return from the S3 sleep state with ACPI enabled (SCI_EN set) and (2) the SCI_EN bit is owned by the hardware and we are not supposed to change it. For this reason, blacklist the affected systems so that the SCI_EN bit is set during resume on them. [NOTE: Unconditional setting SCI_EN for all system on resume doesn't work, because it makes some other systems crash (that's to be expected). Also, it is not entirely clear right now if all of the Apple boxes require this workaround.] This patch fixes the recent regression tracked as http://bugzilla.k

[PATCH] Fix misspellings in pm.h macros

Alan Stern — 2008-11-17T16:14:19

This patch (as1167) fixes some misspellings in various recently-added macros in pm.h. Fortunately these macros are not yet used anywhere. Signed-off-by: Alan Stern rowland.harvard.edu> CC: Rafael J. Wysocki sisk.pl> --- Index: usb-2.6/include/linux/pm.h =================================================================== --- usb-2.6.orig/include/linux/pm.h +++ usb-2.6/include/linux/pm.h < at >< at > -252,7 +252,7 < at >< at > struct dev_pm_ops { #define PM_EVENT_SLEEP(PM_EVENT_SUSPEND | PM_EVENT_HIBERNATE) #define PM_EVENT_USER_SUSPEND(PM_EVENT_USER | PM_EVENT_SUSPEND) #define PM_EVENT_USER_RESUME(PM_EVENT_USER | PM_EVENT_RESUME) -#define PM_EVENT_REMOTE_WAKEUP(PM_EVENT_REMOTE | PM_EVENT_RESUME) +#define PM_EVENT_REMOTE_RESUME(PM_EVENT_REMOTE | PM_EVENT_RESUME) #define PM_EVENT_AUTO_SUSPEND(PM_EVENT_AUTO | PM_EVENT_SUSPEND) #define PM_EVENT_AUTO_RESUME(PM_EVENT_AUTO | PM_EVENT_RESUME) < at >< at > -265,15 +265,15 < at >< at > struct dev_pm_ops { #define PMSG_THAW((struct pm_message){ .event = PM_EVENT_THAW, }) #def

2.6.28-rc5: Reported regressions from 2.6.27

Rafael J. Wysocki — 2008-11-16T16:24:10

[NOTES: (1) Due to some recent controversies, in future I will only use 'Handled-By' tags for people who actually submit a patch or provide substantial help to the reporter (like advising him which commits to revert etc.). (2) We have almost as many regressions with patches as unresolved ones and the majority of these patches have not been merged for over two weeks (for almost three weeks in many cases and for over a _month_ in one case). IMO, this is insane.] This message contains a list of some regressions from 2.6.27, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know. If you know of any other unresolved regressions from 2.6.27, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid. Each entry from the list will be sent additionally in an automatic reply to this message with CCs to the people involved in reporting and handling the issue. List

Suspend regression in 2.6.25.18

Rafael J. Wysocki — 2008-11-16T13:04:22

Hi Thomas, Your commit commit ffa4da2a25bb4ac08f710ac99827baf48a8f8d57 Author: Thomas Gleixner linutronix.de> Date: Wed Sep 3 21:37:08 2008 +0000 clockevents: prevent multiple init/shutdown commit 9c17bcda991000351cb2373f78be7e4b1c44caa3 upstream Signed-off-by: Thomas Gleixner linutronix.de> Signed-off-by: Ingo Molnar elte.hu> Signed-off-by: Greg Kroah-Hartman suse.de> is reported to cause a resume regression on Lenovo x60, x41 and x61 (probably other Lenovo boxes are affected as well) with 2.6.25.18. At the same time, the 2.6.27.y kernels are reported to work correctly for the affected users. Can you please try to identify the updstream commits that fix the problem on these boxes? Thanks, Rafael

pm support

Peter Reid — 2008-11-12T11:30:43

Suspend to disk broken in latest 2.6.28-rc3

Mathieu Chouquet-Stringer — 2008-11-09T19:50:33

Hello Pavel and Rafael, I've just finished compiling a recent snapshot of rc3 (namely gfed4d59 [1]) on a F9 distro (gcc 4.3.0 20080428) and suspend to disk doesn't work anymore. My config is attached below [2]. It just freezes my laptop (an old Thinkpad T30) while it works with kernel version 2.6.26 (I've never tried 2.6.27). I've tried playing with the pm_test facility, and while the "freezer" test works, the "devices" test fails like the real suspend... When that happens, the console still works (ie I can switch VTs and type) but the whole system doesn't respond: I can't login for example... Guess the disk driver is already off at that point. I don't have another computer to capture a serial trace so I guess I'm going to try to bisect it (which is going to take while on this slow computer). If I reconfigure my syslog could I get the output of the suspend process on the screen (for some reason I used to get it during suspend but now I only get it during resume)? Any ideas or pointers to fix this ap

2.6.28-rc3-git6: Reported regressions from 2.6.27

Rafael J. Wysocki — 2008-11-09T17:53:01

This message contains a list of some regressions from 2.6.27, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know. If you know of any other unresolved regressions from 2.6.27, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid. Each entry from the list will be sent additionally in an automatic reply to this message with CCs to the people involved in reporting and handling the issue. Listed regressions statistics: Date Total Pending Unresolved ---------------------------------------- 2008-11-09 73 40 27 2008-11-02 55 41 29 2008-10-25 26 25 20 Unresolved regressions ---------------------- Bug-Entry: http://bugzilla.kernel.org/show_bug.cgi?id=11996 Subject: Tracing framework regression in 2.6.28-rc3 Submitter: Pekka Paalanen iki.fi> Date: 2008-11-09 10:13 (1 days old) Reference

Wakeup events?

Peter Reid — 2008-11-08T14:50:45

[PATCH -next] Hibernate: Do not oops on resume if imagedata are incorrect

Rafael J. Wysocki — 2008-11-08T12:57:42

Hi Len, Please add this patch to the suspend branch. Thanks, Rafael --- From: Rafael J. Wysocki sisk.pl> Subject: Hibernate: Do not oops on resume if image data are incorrect During resume from hibernation using the userland interface image data are being passed from the used space process to the kernel. These data need not be valid, but currently the kernel sometimes oopses if it gets invalid image data, which is wrong. Make the kernel return error codes to the user space in such cases. Signed-off-by: Rafael J. Wysocki sisk.pl> Cc: Pavel Machek suse.cz> --- kernel/power/snapshot.c | 43 ++++++++++++++++++++++++++++++++----------- 1 file changed, 32 insertions(+), 11 deletions(-) Index: linux-2.6/kernel/power/snapshot.c =================================================================== --- linux-2.6.orig/kernel/power/snapshot.c +++ linux-2.6/kernel/power/snapshot.c < at >< at > -529,6 +529,14 < at >< at > static int memory_bm_test_bit(struct mem return test_bit(bit, addr); } +static bool memory

[PATCH -next] Hibernate: Take overlapping zones intoaccount

Rafael J. Wysocki — 2008-11-08T12:56:12

Hi Len, Please add this patch to the suspend branch. Thanks, Rafael --- From: Rafael J. Wysocki sisk.pl> Subject: Hibernate: Take overlapping zones into account It has been requested to make hibernation work with memory hotplugging enabled and for this purpose the hibernation code has to be reworked to take the possible overlapping of zones into account. Thus, rework the hibernation memory bitmaps code to prevent duplication of PFNs from occuring and add checks to make sure that one page frame will not be marked as saveable many times. Additionally, use list.h lists instead of open-coded lists to implement the memory bitmaps. Signed-off-by: Rafael J. Wysocki sisk.pl> Cc: Pavel Machek suse.cz> Cc: Dave Hansen linux.vnet.ibm.com> Cc: Andy Whitcroft shadowen.org> --- kernel/power/snapshot.c | 331 ++++++++++++++++++++++++------------------------ 1 file changed, 169 insertions(+), 162 deletions(-) Index: linux-2.6/kernel/power/snapshot.c =====================================

[PATCH] Fix __pfn_to_page(pfn) for CONFIG_DISCONTIGMEM=y

Rafael J. Wysocki — 2008-11-08T12:53:33

Hi Andrew, This would be good to have in .28 IMO. Thanks, Rafael --- From: Rafael J. Wysocki sisk.pl> Subject: Fix __pfn_to_page(pfn) for CONFIG_DISCONTIGMEM=y Fix macro __pfn_to_page(pfn) so that it doesn't evaluate its argument twice in the CONFIG_DISCONTIGMEM=y case, because 'pfn' may be a result of a funtion call having side effects. For example, the hibernation code applies pfn_to_page(pfn) to the result of a function returning the pfn corresponding to the next set bit in a bitmap and the current bit position is modified on each call. This leads to "interesting" failures for CONFIG_DISCONTIGMEM=y due to the current behavior of __pfn_to_page(pfn). Signed-off-by: Rafael J. Wysocki sisk.pl> Cc: Pavel Machek suse.cz> --- include/asm-generic/memory_model.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) Index: linux-2.6/include/asm-generic/memory_model.h =================================================================== --- linux-2.6.orig/include/asm-generic/memory_model.

Bugs on aspire one A150

Maxim Levitsky — 2008-11-08T03:26:19

I have just bought an Aspire one A150, XP version, as it was the only available here, and installed ubuntu on it. Bugs I discovered so far: ** 1 - embedded controler works in polling mode, due to this: [ 0.708571] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 1.224043] ACPI: EC: missing confirmations, switch off interrupt mode. Maybe this is the reason for the fact that gnome power manager freezes when I unplug the AC, and freezes often when I try to see battery status. (Note: same is seen on my acer aspire 5720) ** 2 - wireless: not to mention the fact that ath5k wasn't installed by default in ubuntu... wireless more or less works, but kernel log is full of backtraces. Was able to connect to my WPA2 access point. Sometimes wireless fails completely, especially after suspend to ram. Advanced features like monitor/injection work, but when I changed the card's mac address it stopped working. I also noticed that if I then start airodump, then wireless works with new mac.

Dynamic Idle?

Peter Reid — 2008-11-07T13:01:24

[Regression] Suspend failure on NForce4-based boards dueto chanes in stop_machine

Rafael J. Wysocki — 2008-11-03T00:28:20

Hi, Current mainline (.28-rc3 at the moment) fails to suspend on at least some machines with dual-core AMD CPUs and NForce4-based mainboards. The affected boxes just hang solid during suspend/hibernation, after suspending devices, probably while the non-boot CPUs are being stopped (I'm able to reproduce this on two different machines). Although this is not reproducible 100% of the time, it is reproducible enough to allow me to carry out bisection, which turned up the following commit as the source of the problem: commit c9583e55fa2b08a230c549bd1e3c0bde6c50d9cc Author: Heiko Carstens de.ibm.com> Date: Mon Oct 13 23:50:10 2008 +0200 stop_machine: use workqueues instead of kernel threads Convert stop_machine to a workqueue based approach. Instead of using kernel threads for stop_machine we now use a an rt workqueue to synchronize all cpus. This has the advantage that all needed per cpu threads are already created when stop_machine gets called. And therefore a call

2.6.28-rc2-git7: Reported regressions from 2.6.27

Rafael J. Wysocki — 2008-11-02T16:04:20

This message contains a list of some regressions from 2.6.27, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know. If you know of any other unresolved regressions from 2.6.27, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid. Each entry from the list will be sent additionally in an automatic reply to this message with CCs to the people involved in reporting and handling the issue. Listed regressions statistics: Date Total Pending Unresolved ---------------------------------------- 2008-11-02 55 41 29 2008-10-25 26 25 20 Unresolved regressions ---------------------- Bug-Entry: http://bugzilla.kernel.org/show_bug.cgi?id=11937 Subject: ext3 __log_wait_for_space: no transactions Submitter: Meelis Roos linux.ee> Date: 2008-10-30 9:49 (4 days old) References: http://marc.info/?l=linux-kernel&m=122

[RFC PATCH 0/2] orinoco: Don't keep cached firmwarearound permanently

David Kilroy — 2008-10-31T01:15:41

The recent patch to load orinoco firmware correctly on resume simply kept the firmware in RAM for the entire time the module was loaded, and only applied to Agere firmware. The first patch uses the new power management notifiers to load and release the firmware prior to suspend and after resume (for both Agere and Symbol). I'm not an expert on where suspend/resume is headed, so I'd appreciate any advice from linux-pm on whether this is the preferred way to do things. The second patch makes the spectrum_cs driver actually invoke this functionality during resume. David Kilroy (2): orinoco: Use PM notifier to cache firmware for use during resume orinoco: Resume spectrum_cs in the same way as orinoco_cs drivers/net/wireless/orinoco.c | 140 ++++++++++++++++++++++++++++-------- drivers/net/wireless/orinoco.h | 8 ++- drivers/net/wireless/spectrum_cs.c | 21 +++++- 3 files changed, 135 insertions(+), 34 deletions(-)