Bug 10724
|
Description
Rafael J. Wysocki
2008-05-16 13:18:43 UTC
Justin said: "Looking at my var/log/* it's been two days and I have not received this message, The only configuration change was disabling irq balancing, changing 300MHZ to 100MHZ, and choosing y to CONFIG_PREEMPT_NONE=y." References : http://lkml.org/lkml/2008/5/18/168 The problem appears to be present in 2.6.26-rc4. References : http://lkml.org/lkml/2008/5/25/195 On Thu, May 29, 2008 at 9:23 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #2 from rjw@sisk.pl 2008-05-29 14:23 ------- > The problem appears to be present in 2.6.26-rc4. > References : http://lkml.org/lkml/2008/5/25/195 > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I'm still receiving this message, I'm currently still using 2.6.26-rc3. I'm planning on updating, but need to do some writing exercise first. can you please do do this test in a working kernel(kernel w/o "ACPI: EC: GPE storm detected, disabling EC GPE" in dmesg)? $cd /sys/firmware/acpi/interrupts $grep . * and attach the result here, this may indicates if there is an actual interrupt storm. Created attachment 16357 [details]
Patch: debug EC GPE
If it's not an interrupt storm, this patch may help us find a more suitable threshold for detecting EC GPE storm.
Justin, please apply the debug patch and attach the dmesg output.
On Mon, Jun 2, 2008 at 2:04 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #5 from rui.zhang@intel.com 2008-06-01 19:04 ------- > Created an attachment (id=16357) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16357&action=view) > Patch: debug EC GPE > > If it's not an interrupt storm, this patch may help us find a more suitable > threshold for detecting EC GPE storm. > Justin, please apply the debug patch and attach the dmesg output. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. I recompiled attached is dmesg I'm seeing a: [ 115.021943] ACPI: EC: 6 EC interrupts detected, EC GPE storm? [ 115.021943] ACPI: EC: 7 EC interrupts detected, EC GPE storm? [ 115.026666] ACPI: EC: 8 EC interrupts detected, EC GPE storm? [ 115.036630] ACPI: EC: 9 EC interrupts detected, EC GPE storm? Which probably means something not good!! regards; and thank you for the help. could you please attach the test result I asked in comment #4 please? Created attachment 16362 [details]
ACPI:EC:GPE storm detected, disabling EC GPE
Hope this help's with this issue.
On Mon, Jun 2, 2008 at 5:32 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > rui.zhang@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Status|ASSIGNED |NEEDINFO > > > > > ------- Comment #7 from rui.zhang@intel.com 2008-06-01 22:32 ------- > could you please attach the test result I asked in comment #4 please? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. . Hope this helps. regards; Hah, valuable info. thanks. This problem is caused by commit fa95ba04e6ba11d71e1b87becd054b38faf546c8. I don't know if we should call it a bug or not as EC poll mode should still work for you, right? Justin, do you get any other problems, like something failed to work anymore, after seeing the " ACPI: EC: GPE storm detected, disabling EC GPE" message? Alexey, maybe it's too aggressive to switch to poll mode if more than 5 EC GPEs are not handled in time? On Mon, Jun 2, 2008 at 6:29 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #10 from rui.zhang@intel.com 2008-06-01 23:29 ------- > Hah, valuable info. thanks. > > This problem is caused by commit fa95ba04e6ba11d71e1b87becd054b38faf546c8. > I don't know if we should call it a bug or not as EC poll mode should still > work for you, right? > > Justin, > do you get any other problems, like something failed to work anymore, after > seeing the " ACPI: EC: GPE storm detected, disabling EC GPE" message? > > Alexey, > maybe it's too aggressive to switch to poll mode if more than 5 EC GPEs are > not handled in time? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > No everything runs fine, The only weird thing I see is with ift-load and uvcvideo from time to time it will not load when I start the system up.(usually after I recompile the kernel). regards; Rui, 5 was a wild guess. In theory, we should not get even 2 if the EC is working right. Now it seems that almost all cheap machines are shipped with broken EC and not all the people are happy with the switch to poll mode on these machines. Were could be two solutions -- one is to increase cut-off to say 10, and the other to change from udelay(ACPI_EC_DELAY) to msleep(1). Created attachment 16364 [details]
Use msleep instead of udelay
Please test this patch first
Created attachment 16365 [details]
Increase threshold for stray interrupts
Please also test this one.
On Mon, Jun 2, 2008 at 7:41 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #14 from astarikovskiy@suse.de 2008-06-02 00:41 ------- > Created an attachment (id=16365) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view) > Increase threshold for stray interrupts > > Please also test this one. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Sorry for keeping you waiting, I'm over on the west coast(Socal), I'll give these a try and see what happens regards; On Mon, Jun 2, 2008 at 4:25 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Mon, Jun 2, 2008 at 7:41 AM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> >> >> >> ------- Comment #14 from astarikovskiy@suse.de 2008-06-02 00:41 ------- >> Created an attachment (id=16365) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view) >> Increase threshold for stray interrupts >> >> Please also test this one. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > Sorry for keeping you waiting, I'm over on the west coast(Socal), I'll > give these a try and see what happens > regards; > > -- > Justin P. Mattock > Alright I applied the patches, and am not seeing anything of this nature("knock on wood"), I'll let the system run as is for a while(1hrs or 2hrs) and post dmesg. Hopefully this takes care of it. regards; On Mon, Jun 2, 2008 at 5:18 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Mon, Jun 2, 2008 at 4:25 PM, Justin Mattock <justinmattock@gmail.com> > wrote: >> On Mon, Jun 2, 2008 at 7:41 AM, <bugme-daemon@bugzilla.kernel.org> wrote: >>> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >>> >>> >>> >>> >>> >>> ------- Comment #14 from astarikovskiy@suse.de 2008-06-02 00:41 ------- >>> Created an attachment (id=16365) >>> --> (http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view) >>> Increase threshold for stray interrupts >>> >>> Please also test this one. >>> >>> >>> -- >>> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >>> ------- You are receiving this mail because: ------- >>> You are on the CC list for the bug, or are watching someone who is. >>> >> >> Sorry for keeping you waiting, I'm over on the west coast(Socal), I'll >> give these a try and see what happens >> regards; >> >> -- >> Justin P. Mattock >> > > Alright I applied the patches, and am not seeing anything of this > nature("knock on wood"), I'll let the system run as is for > a while(1hrs or 2hrs) and post dmesg. Hopefully this takes care of it. > regards; > > -- > Justin P. Mattock > O.K. here is dmesg of the system running after around an hour; regards; Regressions list annotation: Handled-by: Alexey Starikovskiy <astarikovskiy@suse.de> Patch : http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view Patch : http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view On Thu, Jun 5, 2008 at 9:13 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #18 from rjw@sisk.pl 2008-06-05 14:13 ------- > Regressions list annotation: > Handled-by: Alexey Starikovskiy <astarikovskiy@suse.de> > Patch : http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view > Patch : http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > after seeing a post about changing 5 to 20 I decided to give this a try: So after an hour I of having the system running with the change I have not seen this message. Keep in mind I did patch the file, but after a few days I decide to put everything back and start fresh. So the only changes is this below: drivers/acpi/ec.c @@ -527,47 +488,51 @@ static u32 acpi_ec_gpe_handler(void *data) { acpi_status status = AE_OK; struct acpi_ec *ec = data; u8 state = acpi_ec_read_status(ec); pr_debug(PREFIX "~~~> interrupt\n"); atomic_inc(&ec->irq_count); - if (atomic_read(&ec->irq_count) > 5) { + if (atomic_read(&ec->irq_count) > 20) { pr_err(PREFIX "GPE storm detected, disabling EC GPE\n"); ec_switch_to_poll_mode(ec); goto end; } hopefully this works. regards; msleep patch in comment #13 added to acpi-test tree Created attachment 16447 [details]
Disable GPE during transaction only
Please check that this patch helps.
On Tue, Jun 10, 2008 at 1:54 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > astarikovskiy@suse.de changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Attachment #16364 [details]|0 |1 > is obsolete| | > Attachment #16365 [details]|0 |1 > is obsolete| | > AssignedTo|rui.zhang@intel.com |astarikovskiy@suse.de > Status|REOPENED |ASSIGNED > > > > > ------- Comment #21 from astarikovskiy@suse.de 2008-06-10 06:54 ------- > Created an attachment (id=16447) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16447&action=view) > Disable GPE during transaction only > > Please check that this patch helps. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Will do, I'll let you know of the status. regards; On Tue, Jun 10, 2008 at 2:01 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Tue, Jun 10, 2008 at 1:54 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> astarikovskiy@suse.de changed: >> >> What |Removed |Added >> ---------------------------------------------------------------------------- >> Attachment #16364 [details]|0 |1 >> is obsolete| | >> Attachment #16365 [details]|0 |1 >> is obsolete| | >> AssignedTo|rui.zhang@intel.com |astarikovskiy@suse.de >> Status|REOPENED |ASSIGNED >> >> >> >> >> ------- Comment #21 from astarikovskiy@suse.de 2008-06-10 06:54 ------- >> Created an attachment (id=16447) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=16447&action=view) >> Disable GPE during transaction only >> >> Please check that this patch helps. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > Will do, I'll let you know of the status. > regards; > > -- > Justin P. Mattock > Sorry for the latency,(I stepped out for a while), I wanted to make sure I have this correct with the patch that you had sent me today do I just apply that one to a fresh copy of ec.c or do I need to apply today's, plus the other two that was sent to me last week? I went ahead and applied the patch from today but only today on a fresh copy of ec.c, I did notice the gpe storm message, as well as Hal having difficulty loading, as well as funkiness from the battery going from plugged-in to unplugged not showing any info in dmesg. from that I was skeptical with keeping that running so I went and put everything back just to be safe. regards; Justin, battery and AC drivers are read-only to HW. They do not change the state of HW in any way, thus it is not very critical to allow them being "broken" or oddly-behaving. regards. On Tue, Jun 10, 2008 at 9:30 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #24 from astarikovskiy@suse.de 2008-06-10 14:30 ------- > Justin, > battery and AC drivers are read-only to HW. They do not change the state of > HW > in any way, thus it is not very critical to allow them being "broken" or > oddly-behaving. > regards. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K.,("Strange behavior from over here, for some reason") Anyways As for applying the patches I am willing to try, do I need to apply the three to ec.c or just today's, to a fresh copy of ec.c, Also just to let you know that on line 191 of ec.c with today's patch applied this need's to be removed ec_schedule_ec_poll(ec); as well, so you can compile,(I realize people get tired and miss things here and there, which is O.K.); regards; Justin, what do you see when you apply only the patch from comment #13? If it functions properly, do you notice any performance difference, say, when accessing battery files? On Wed, Jun 11, 2008 at 4:11 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > len.brown@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Status|ASSIGNED |NEEDINFO > > > > > ------- Comment #26 from len.brown@intel.com 2008-06-10 21:11 ------- > Justin, > what do you see when you apply only the patch from comment #13? > If it functions properly, do you notice any performance difference, > say, when accessing battery files? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Hello; It's midnight right now so I'll have to give you a better response tomorrow; anyways to give you as much info as possible: As a small test I wanted to see how the system would react without the battery, and yes the message seems to not be there. i.g. As a small test running the system without the battery has given me 5+ hours without this message.(leading me to believe this has to do with the battery or something in the same vicinity). I think whats happening is the people with the other laptop's are noticing key's and such, as for people with mac's there noticing battery issues. regards; The patch from comment #13 shipped in linux-2.6.26-rc6, please let us know if that works or if we still have an issue with this box. commit 1b7fc5aae8867046f8d3d45808309d5b7f2e036a Author: Alexey Starikovskiy <astarikovskiy@suse.de> Date: Fri Jun 6 11:49:33 2008 -0400 ACPI: EC: Use msleep instead of udelay while waiting for event. On Fri, Jun 13, 2008 at 5:17 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > len.brown@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Status|NEEDINFO |RESOLVED > Resolution| |CODE_FIX > > > > > ------- Comment #28 from len.brown@intel.com 2008-06-12 22:17 ------- > The patch from comment #13 shipped in linux-2.6.26-rc6, > please let us know if that works or if we still > have an issue with this box. > > commit 1b7fc5aae8867046f8d3d45808309d5b7f2e036a > Author: Alexey Starikovskiy <astarikovskiy@suse.de> > Date: Fri Jun 6 11:49:33 2008 -0400 > > ACPI: EC: Use msleep instead of udelay while waiting for event. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. didn't mean to keep you waiting, I'll give it a try. regards; On Fri, Jun 13, 2008 at 6:23 AM, Justin Mattock <justinmattock@gmail.com> wrote: > On Fri, Jun 13, 2008 at 5:17 AM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> len.brown@intel.com changed: >> >> What |Removed |Added >> ---------------------------------------------------------------------------- >> Status|NEEDINFO |RESOLVED >> Resolution| |CODE_FIX >> >> >> >> >> ------- Comment #28 from len.brown@intel.com 2008-06-12 22:17 ------- >> The patch from comment #13 shipped in linux-2.6.26-rc6, >> please let us know if that works or if we still >> have an issue with this box. >> >> commit 1b7fc5aae8867046f8d3d45808309d5b7f2e036a >> Author: Alexey Starikovskiy <astarikovskiy@suse.de> >> Date: Fri Jun 6 11:49:33 2008 -0400 >> >> ACPI: EC: Use msleep instead of udelay while waiting for event. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > O.K. didn't mean to keep you waiting, I'll give it a try. > regards; > > -- > Justin P. Mattock > Ooops I didn't understand you're message, "know wonder I saw it in there", O.K. anyways, after being up all day today(pulled this morning) which has msleep in there, I'm still seeing this message. argv(I really don't know how or what to do with this one"). for the time being I'll just keep my battery removed, so I don't see this message. regards; Hi all, I cannot confirm that 2.6.26-rc6 fixes that. I have seen this in all the -rc series but not 2.6.25. Now I still get: ACPI: EC: non-query interrupt received, switching to interrupt mode ACPI: EC: GPE storm detected, disabling EC GPE which I didn't have in 2.6.25. 2.6.26-rc6 #1 SMP PREEMPT Acer TravelMate 3012 Doing the grep . /sys/firmware/acpi/interrupts/* stuff I see all zeros around, but still: gpe_all:6 sci:6 so that are the only lines with a non-zero value. (all with 2.6.26-rc6) Is there more I can provide? Bye Norbert Message "ACPI: EC: GPE storm detected, disabling EC GPE" should appear on all broken hardware, we are not trying to "fix" that. Do you see any _problems_, like battery/ac state not updating, etc? On Fri, Jun 13, 2008 at 8:53 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #32 from astarikovskiy@suse.de 2008-06-13 01:53 ------- > Message "ACPI: EC: GPE storm detected, disabling EC GPE" should appear on all > broken hardware, we are not trying to "fix" that. > Do you see any _problems_, like battery/ac state not updating, etc? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > yes, when disconnecting the A/C adapter after the message appears, issuing acpi in a terminal results with nothing on how much time left etc.., and also if this message appears, and I disconnect the battery itself, and reattach I will not see any info on the battery, normally I do. I think here with the mac's It something with the battery, and not power buttons, or keyboard keys(but could be wrong). regards please check the contents of /proc/acpi/ac_adapter and /proc/acpi/battery/*/ for these situations. As I understand state of ac_adapter does not change as you plug/un-plug it, right? Same with the battery, yes? On Fri, Jun 13, 2008 at 10:03 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #34 from astarikovskiy@suse.de 2008-06-13 15:03 ------- > please check the contents of /proc/acpi/ac_adapter and /proc/acpi/battery/*/ > for these situations. As I understand state of ac_adapter does not change as > you plug/un-plug it, right? Same with the battery, yes? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Usually when I disconect the battery and reconect there will be a message in dmesg attached is when I connect the battery, when the message appears this will not be seen. I'll have a check in the files also /proc/acpi/* regards Created attachment 16474 [details]
Disable GPE while query is pending
Please check if this patch works better (on top of -rc6)
On Sat, Jun 14, 2008 at 12:11 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > astarikovskiy@suse.de changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Attachment #16447 [details]|0 |1 > is obsolete| | > > > > > ------- Comment #36 from astarikovskiy@suse.de 2008-06-14 05:11 ------- > Created an attachment (id=16474) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16474&action=view) > Disable GPE while query is pending > > Please check if this patch works better (on top of -rc6) > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Sure, I might of tried this patch already, but that was for rc5, hopefully this fixes this. On Sat, Jun 14, 2008 at 8:16 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Sat, Jun 14, 2008 at 12:11 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> astarikovskiy@suse.de changed: >> >> What |Removed |Added >> ---------------------------------------------------------------------------- >> Attachment #16447 [details]|0 |1 >> is obsolete| | >> >> >> >> >> ------- Comment #36 from astarikovskiy@suse.de 2008-06-14 05:11 ------- >> Created an attachment (id=16474) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=16474&action=view) >> Disable GPE while query is pending >> >> Please check if this patch works better (on top of -rc6) >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > Sure, I might of tried this patch already, but that was for rc5, > hopefully this fixes this. > > -- > Justin P. Mattock > O.K. I pulled the latest git, and applied the patch, on the first try, I made the mistake of changing 5 to 20, i.g. if (atomic_read(&ec->irq_count) > 5 &&, So after putting 20 back to 5, with everything as they should I see this in dmesg: [ 208.352562] ACPI: EC: GPE storm detected, disabling EC GPE [ 208.454844] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 208.532300] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 208.599015] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 244.123340] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 275.941465] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 307.755635] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 339.450494] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 371.562684] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 403.161222] ACPI: EC: non-query interrupt received, switching to interrupt mode [ 408.460630] applesmc: wait status failed: 5 != 4 [ 408.723819] applesmc: wait status failed: c != 8 [ 434.952505] ACPI: EC: non-query interrupt received, switching to interrupt mode maybe changing 5 to 20 is not a bad idea, but then again maybe it is due to not seeing this message right away. I'm going to try putting 5 to 20, just to see. Justin, I'm interested in ac_adapter state correctness with the vanilla kernel + last patch alone. You could set 5 to 20 or 100 or even delete the whole detection afterwards. Thanks. On Sun, Jun 15, 2008 at 5:35 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #39 from astarikovskiy@suse.de 2008-06-15 10:35 ------- > Justin, I'm interested in ac_adapter state correctness with the vanilla > kernel > + last patch alone. You could set 5 to 20 or 100 or even delete the whole > detection afterwards. Thanks. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. after setting to 5, you saw what I was seeing, I've set this to 20 now, to see what happens. As a status: yesterday the system was up for 2:38:00 hrs without this message. So today I'm going to try and leave the system running all day to see. The last time I did this with udelay set in ec.c the message appeared after three hours., So maybe msleep will be different. regards; On Sun, Jun 15, 2008 at 5:47 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Sun, Jun 15, 2008 at 5:35 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> >> >> >> ------- Comment #39 from astarikovskiy@suse.de 2008-06-15 10:35 ------- >> Justin, I'm interested in ac_adapter state correctness with the vanilla >> kernel >> + last patch alone. You could set 5 to 20 or 100 or even delete the whole >> detection afterwards. Thanks. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > O.K. after setting to 5, you saw what I was seeing, I've set this to > 20 now, to see what happens. As a status: > yesterday the system was up for 2:38:00 hrs without this message. So > today I'm going to try and leave the system running all day to see. > The last time I did this with udelay set in ec.c the message appeared > after three hours., So maybe msleep will be different. > regards; > -- > Justin P. Mattock > Attached is dmesg of the system being up for a while, after around five hours the GPE storm is triggered: So if I set 5 to 20 then to 100 will that hurt anything on the system? On Sun, Jun 15, 2008 at 11:00 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Sun, Jun 15, 2008 at 5:47 PM, Justin Mattock <justinmattock@gmail.com> > wrote: >> On Sun, Jun 15, 2008 at 5:35 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >>> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >>> >>> >>> >>> >>> >>> ------- Comment #39 from astarikovskiy@suse.de 2008-06-15 10:35 ------- >>> Justin, I'm interested in ac_adapter state correctness with the vanilla >>> kernel >>> + last patch alone. You could set 5 to 20 or 100 or even delete the whole >>> detection afterwards. Thanks. >>> >>> >>> -- >>> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >>> ------- You are receiving this mail because: ------- >>> You are on the CC list for the bug, or are watching someone who is. >>> >> >> O.K. after setting to 5, you saw what I was seeing, I've set this to >> 20 now, to see what happens. As a status: >> yesterday the system was up for 2:38:00 hrs without this message. So >> today I'm going to try and leave the system running all day to see. >> The last time I did this with udelay set in ec.c the message appeared >> after three hours., So maybe msleep will be different. >> regards; >> -- >> Justin P. Mattock >> > > Attached is dmesg of the system being up for a while, after around > five hours the GPE storm is triggered: > So if I set 5 to 20 then to 100 will that hurt anything on the system? > > -- > Justin P. Mattock > Also the state of the A/C adapter is real sketchy after the message is triggered, when unplugging the A/C, the dimmer for the screen does not trigger, on the first or second disconnect, but eventually will, upon plugging the A/C the dimmer will not brighten the screen on some tries, then I noticed pommed was becoming reversed. i.g. unplugging, the screen does not dim, upon plugging in the A/C the screen dims. There is info on the battery state ans so forth in dmesg. I guess setting 20 to 100 will take longer for the GPE storm message to be triggered, which is good in my case, due to not having my system on for long periods of time, but not so good for servers and such that are on for weeks or month's at a time. On Sunday, 22 of June 2008, Justin Mattock wrote:
> On Sun, Jun 22, 2008 at 5:54 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> > Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> > Submitter : Justin Mattock <justinmattock@gmail.com>
> > Date : 2008-05-16 6:17 (38 days old)
> > References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> > http://lkml.org/lkml/2008/5/18/168
> > http://lkml.org/lkml/2008/5/25/195
> > Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> >
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
> >
> >
> >
>
> Yes; setting if (atomic_read(&ec->irq_count) > 5) { to
> if (atomic_read(&ec->irq_count) > 20) {
> in drivers/acpi/ac.c only prolongs the bug, after a period of time(3
> to 5 hrs) the the message does appear.
setting to UNRESOLVED since per comment #28, as we shipped the msleep(1) patch from comment #13 and we still have a problem in 2.6.26-rc8. (should we revert that patch?) and various incantations of testing the patch in comment #14 have apparently not led to a conclusion. On Thu, Jun 26, 2008 at 5:12 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > len.brown@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Status|RESOLVED |REOPENED > Resolution|CODE_FIX | > > > > > ------- Comment #44 from len.brown@intel.com 2008-06-25 22:12 ------- > setting to UNRESOLVED > since per comment #28, as we shipped the msleep(1) patch > from comment #13 and we still have a problem in 2.6.26-rc8. > > (should we revert that patch?) > > and various incantations of testing the patch in comment #14 > have apparently not led to a conclusion. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > If you would like to, I don't see a problem with that. Does a 2.6.26-rc8 kernel that is both unpatched and without any external modules loaded exhibit this problem? If yes, please attach the output of "dmesg -s 1000000". On Fri, Jun 27, 2008 at 9:42 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > bunk@kernel.org changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |bunk@kernel.org > > > > > ------- Comment #46 from bunk@kernel.org 2008-06-27 14:42 ------- > Does a 2.6.26-rc8 kernel that is both unpatched and without any external > modules loaded exhibit this problem? > > If yes, please attach the output of "dmesg -s 1000000". > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > let me compile a fresh kernel and give you the results(dmesg) So give me some time for this as well as the others. regards; On Fri, Jun 27, 2008 at 10:15 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Fri, Jun 27, 2008 at 9:42 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> bunk@kernel.org changed: >> >> What |Removed |Added >> ---------------------------------------------------------------------------- >> CC| |bunk@kernel.org >> >> >> >> >> ------- Comment #46 from bunk@kernel.org 2008-06-27 14:42 ------- >> Does a 2.6.26-rc8 kernel that is both unpatched and without any external >> modules loaded exhibit this problem? >> >> If yes, please attach the output of "dmesg -s 1000000". >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > let me compile a fresh kernel and give you the results(dmesg) > So give me some time for this as well as the others. > regards; > > -- > Justin P. Mattock > O.K. I loaded a fresh kernel hot off the oven: attached is dmesg1 after around 13min the message appears, and dmesg2 of the message going off during boot. Keep in mind there are no external modules installed. regards (In reply to comment #48) > On Fri, Jun 27, 2008 at 10:15 PM, Justin Mattock > O.K. I loaded a fresh kernel hot off the oven: attached is dmesg1 > after around 13min the message appears, and dmesg2 > of the message going off during boot. It seems you sent them as email attachments, which means I'm not getting them. Please attach them to the bug. Thanks On Sun, Jun 29, 2008 at 6:21 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #49 from bunk@kernel.org 2008-06-29 11:21 ------- > (In reply to comment #48) >> On Fri, Jun 27, 2008 at 10:15 PM, Justin Mattock > >> O.K. I loaded a fresh kernel hot off the oven: attached is dmesg1 >> after around 13min the message appears, and dmesg2 >> of the message going off during boot. > > It seems you sent them as email attachments, which means I'm not getting > them. > > Please attach them to the bug. > > Thanks > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. I'll do that right away. regards; Created attachment 16651 [details]
no external modules
here is dmesg of gpe storm after 13min of the system being up.
there are no external modules at all.
Created attachment 16652 [details]
gpe storm during boot
this was right away during a reboot.
Created attachment 16653 [details]
remove A/C adapter
removing A/C adapter show's the GPE storm message
(In reply to comment #39) > Justin, I'm interested in ac_adapter state correctness with the vanilla > kernel > + last patch alone. You could set 5 to 20 or 100 or even delete the whole > detection afterwards. Thanks. > Hello; after a week of getting carried away with other parts of the kernel, I decided to see what the reaction is with unplugging and plugging in the A/C adapter during a gpe storm. With 2.6.26-rc7 it seemed very inconsistent, now with 2.6.26-rc8 I'm seeing right reactions with pommed(i.g. unplugged: screen goes dim, battery info in dmesg etc..) Aside from this everything seems O.K. Justin, Once again, EC in your machine is not conforming to spec, it sends too many interrupts during its activity. Detection for this is new in 2.6.26-rc, so you are expected to see "GPE storm detected" and this is not a bug in Linux, rather it is bug in your machine (which we can only detect and workaround). The only thing we could correct, is that workaround does not make your system less usable than it was without it. So, if there is nothing to complain about beside "GPE storm detected" message, your system works as expected, and there is no bug. On Mon, Jun 30, 2008 at 6:33 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #55 from astarikovskiy@suse.de 2008-06-29 23:33 ------- > Justin, > Once again, EC in your machine is not conforming to spec, it sends too many > interrupts during its activity. Detection for this is new in 2.6.26-rc, so > you > are expected to see "GPE storm detected" and this is not a bug in Linux, > rather > it is bug in your machine (which we can only detect and workaround). > The only thing we could correct, is that workaround does not make your system > less usable than it was without it. So, if there is nothing to complain about > beside "GPE storm detected" message, your system works as expected, and there > is no bug. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Alright, I'm not trying to complain, just trying to see if I can solve this.(or at least try and help). Anyways, If you're saying this is not a bug, then I'm happy with having this closed. As for the workaround; I guess I've been using that(patch) for sometime now. regards; On Sunday, 29 of June 2008, Justin Mattock wrote:
> On Sun, Jun 29, 2008 at 6:18 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> > Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> > Submitter : Justin Mattock <justinmattock@gmail.com>
> > Date : 2008-05-16 6:17 (45 days old)
> > References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> > http://lkml.org/lkml/2008/5/18/168
> > http://lkml.org/lkml/2008/5/25/195
> > Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> >
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
> >
> >
> >
>
> yes; still having this message show up. with a patched kernel and
> external modules, and with
> a fresh kernel with no external modules, or patches.
Created attachment 16735 [details]
adding acpi_osi=Linux
receiving large amounts of interrupts in /sys/firmware/acpi/interrupts/
when using this boot param.
Created attachment 16736 [details]
acpi_osi=Darwin
when using acpi_osi=Darwin I'm receiving a lot less of interrupts
in /sys/firmware/acpi/interrupts/ will monitor all day to
see if this option causes an a detected gpe storm.
On Sunday, 6 of July 2008, Justin Mattock wrote:
> On Sun, Jul 6, 2008 at 11:45 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> > Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> > Submitter : Justin Mattock <justinmattock@gmail.com>
> > Date : 2008-05-16 6:17 (52 days old)
> > References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> > http://lkml.org/lkml/2008/5/18/168
> > http://lkml.org/lkml/2008/5/25/195
> > Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> >
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
> >
> >
> >
>
> yes;
> From what I can see so far, the reason for the gpe storm detector going off
> is due to too many interrupts with the battery. As an example there was
> a bug filed for the same issue with the macbook's:
> https://bugzilla.novell.com/show_bug.cgi?id=301365#c10
> using acpi_osi=Darwin does prevent the gpe storm detector from
> going off, but you loose any info on the battery. If anybody has any ideas on
> modifying any of the battery modules, or adjusting the DSDT it sure
> would be appreciated.
> regards;
please attach the acpidump. Please attach the result of "grep . /sys/firmware/acpi/interrupts/*" for both acpi_osi=Linux and acpi_osi=Darwin Created attachment 16751 [details]
acpidump
acpidump of macbook2,2
Created attachment 16752 [details]
acpi_osi=Linux
copy of /sys/firmware/acpi/interrupts/*
using acpi_osi=Linux
Created attachment 16753 [details]
acpi_osi=Darwin results of /sys/firmware/acpi/interrupts/*
copy of /sys/firmware/acpi/interrupts/*
when using acpi_osi=Darwin boot option
Created attachment 16754 [details]
dsdt.dsl macbook2,2
copy of dsdt.dsl "just for references"
On Mon, Jul 7, 2008 at 5:27 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #61 from rui.zhang@intel.com 2008-07-06 22:26 ------- > please attach the acpidump. > Please attach the result of "grep . /sys/firmware/acpi/interrupts/*" for both > acpi_osi=Linux and acpi_osi=Darwin > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. here you go, let me know if you need anything else. regards; Created attachment 16759 [details]
Register SBS notify callback
Please check if this patch helps to load sbshc/sbs drivers.
On Mon, Jul 7, 2008 at 12:24 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #67 from astarikovskiy@suse.de 2008-07-07 05:24 ------- > Created an attachment (id=16759) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16759&action=view) > Register SBS notify callback > > Please check if this patch helps to load sbshc/sbs drivers. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.K. I'll let you know if it loads. Created attachment 16764 [details]
sbs module with patch
here is dmesg of the (patched)sbs module being loaded after booting up.
unfortunately loading the module during boot causes the system to hang, with
the acpi_osi=Dawrin option.
On Mon, Jul 7, 2008 at 4:32 PM, Justin Mattock <justinmattock@gmail.com> wrote: > On Mon, Jul 7, 2008 at 12:24 PM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> >> >> >> ------- Comment #67 from astarikovskiy@suse.de 2008-07-07 05:24 ------- >> Created an attachment (id=16759) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=16759&action=view) >> Register SBS notify callback >> >> Please check if this patch helps to load sbshc/sbs drivers. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > O.K. I'll let you know if it loads. > > -- > Justin P. Mattock > O.k. I applied the patch, in the bug report is an attachment of dmesg, with the sbs module being loaded. unfortunately still having sbs hang during boot. Does it load if you disable detection of GPE storm? e.g. put 100 instead of 5? On Mon, Jul 7, 2008 at 6:02 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #71 from astarikovskiy@suse.de 2008-07-07 11:02 ------- > Does it load if you disable detection of GPE storm? e.g. put 100 instead of > 5? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I went and took the whole detector out of ec.c to see what would happen, and yes still the same results. then put the detector back in. On Sunday, 13 of July 2008, Justin Mattock wrote:
> On Sun, Jul 13, 2008 at 6:00 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > This message has been generated automatically as a part of a report
> > of recent regressions.
> >
> > The following bug entry is on the current list of known regressions
> > from 2.6.25. Please verify if it still should be listed.
> >
> >
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> > Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> > Submitter : Justin Mattock <justinmattock@gmail.com>
> > Date : 2008-05-16 6:17 (59 days old)
> > References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> > http://lkml.org/lkml/2008/5/18/168
> > http://lkml.org/lkml/2008/5/25/195
> > Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> >
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
> >
> >
> >
>
> Yeah I think it would be a good idea to still have this as a bug.
On Sunday, 13 of July 2008, Vegard Nossum wrote:
> On Sun, Jul 13, 2008 at 9:34 PM, Justin Mattock <justinmattock@gmail.com>
> wrote:
> > On Sun, Jul 13, 2008 at 6:00 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >> This message has been generated automatically as a part of a report
> >> of recent regressions.
> >>
> >> The following bug entry is on the current list of known regressions
> >> from 2.6.25. Please verify if it still should be listed.
> >>
> >>
> >> Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> >> Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> >> Submitter : Justin Mattock <justinmattock@gmail.com>
> >> Date : 2008-05-16 6:17 (59 days old)
> >> References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> >> http://lkml.org/lkml/2008/5/18/168
> >> http://lkml.org/lkml/2008/5/25/195
> >> Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> >>
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
>
> I didn't follow the discussion, but I may contribute the following
> information:
>
> This message first appears in my logs on May 16. That was with kernel
> version 2.6.24.5-85.fc8. The kernel I used before that was
> 2.6.24.4-64.fc8 (May 3). My logs go back to November 8
> (2.6.23.1-42.fc8). So we can hardly consider this a regression since
> 2.6.25, but rather one since 2.6.24?
>
> (I'll also note that this message appears quite infrequently here.
> Only 42 times in 219 boot-ups. So it would be hard to bisect, but I'm
> guessing the error was introduced somewhere between 2.6.24.4 and
> 2.6.24.5.)
On Sunday, 13 of July 2008, Hugh Dickins wrote:
> On Sun, 13 Jul 2008, Vegard Nossum wrote:
> > On Sun, Jul 13, 2008 at 9:34 PM, Justin Mattock <justinmattock@gmail.com>
> wrote:
> > > On Sun, Jul 13, 2008 at 6:00 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > >> This message has been generated automatically as a part of a report
> > >> of recent regressions.
> > >>
> > >> The following bug entry is on the current list of known regressions
> > >> from 2.6.25. Please verify if it still should be listed.
> > >>
> > >>
> > >> Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=10724
> > >> Subject : ACPI: EC: GPE storm detected, disabling EC GPE
> > >> Submitter : Justin Mattock <justinmattock@gmail.com>
> > >> Date : 2008-05-16 6:17 (59 days old)
> > >> References : http://marc.info/?l=linux-kernel&m=121091875711824&w=4
> > >> http://lkml.org/lkml/2008/5/18/168
> > >> http://lkml.org/lkml/2008/5/25/195
> > >> Patch :
> http://bugzilla.kernel.org/attachment.cgi?id=16364&action=view
> > >>
> http://bugzilla.kernel.org/attachment.cgi?id=16365&action=view
> >
> > I didn't follow the discussion, but I may contribute the following
> information:
> >
> > This message first appears in my logs on May 16. That was with kernel
> > version 2.6.24.5-85.fc8. The kernel I used before that was
> > 2.6.24.4-64.fc8 (May 3). My logs go back to November 8
> > (2.6.23.1-42.fc8). So we can hardly consider this a regression since
> > 2.6.25, but rather one since 2.6.24?
> >
> > (I'll also note that this message appears quite infrequently here.
> > Only 42 times in 219 boot-ups. So it would be hard to bisect, but I'm
> > guessing the error was introduced somewhere between 2.6.24.4 and
> > 2.6.24.5.)
>
> You're comparing against Fedora kernels, which often contain
> patches which haven't got into mainline yet. As in this case.
> Unless it used to be assembled from separate pieces, there was
> no "GPE storm detected" message in 2.6.24.N or 2.6.25.N: it was
> added in 2.6.26-rc1.
>
> I sometimes see it too, on a Fujitsu-Siemens laptop.
Justin, Darwin SBS issue seems to be big enough for its own bug report. Vegard, the original problem with this bug was that after EC GPE storm was detected (message...), system became less responsive to AC or special keys. On Sun, Jul 13, 2008 at 9:57 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #76 from astarikovskiy@suse.de 2008-07-13 14:57 ------- > Justin, Darwin SBS issue seems to be big enough for its own bug report. > Vegard, the original problem with this bug was that after EC GPE storm was > detected (message...), system became less responsive to AC or special keys. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Yep; I can agree with that. Created attachment 16862 [details]
Avoid dropping rapid hotkey events (or other GPEs) on Asus EeePC
Please check if this patch helps with all the issues...
On Thu, Jul 17, 2008 at 11:54 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #78 from astarikovskiy@suse.de 2008-07-17 04:54 ------- > Created an attachment (id=16862) > --> (http://bugzilla.kernel.org/attachment.cgi?id=16862&action=view) > Avoid dropping rapid hotkey events (or other GPEs) on Asus EeePC > > Please check if this patch helps with all the issues... > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Cool thanks. The patch 1b7fc5aae8867046f8d3d45808309d5b7f2e036a introduced to fix this bug makes bug http://bugzilla.kernel.org/show_bug.cgi?id=10482 reappear for me, which is fixed if I revert 1b7fc5. On Sun, Jul 20, 2008 at 1:18 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > thomas@archlinux.org changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |thomas@archlinux.org > > > > > ------- Comment #80 from thomas@archlinux.org 2008-07-20 13:18 ------- > The patch 1b7fc5aae8867046f8d3d45808309d5b7f2e036a introduced to fix this bug > makes bug http://bugzilla.kernel.org/show_bug.cgi?id=10482 reappear for me, > which is fixed if I revert 1b7fc5. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > At this moment I'm not sure what is causing the amount of interrupts to the battery mechanism!! i.g. could pommed or another daemon, be causing the vast amount of wake ups or something else. "checking my system for that"; regards; Hello, i googled a bit and found this entry, not sure whether i should post here. I have a similar problem on an Acer Notebook. It's an Acer Aspire 5520 G. I tested some different Kernel versions - 2.6.25.8 is working fine without any problems, although i miss some features of 2.6.26.1 i would really want to use - Anyway, _every_ Kernel version in 2.6.26.* isn't working - The Kernel seems to end in a loop or the Kernel is waiting for something after showing: ACPI: EC: GPE storm detected, disabling EC GPE Anyway, let's give some other _maybe_ helpful information on this: Booting up with framebuffer (vga=773 ... video=vesafb:mtrr:4,ywrap) gives a blank screen - if i press the "power" button on my notebook the screen appears and i see some acpi messages related to pressing the power button. Booting without framebuffer gives no blank screen (vga=normal) the last message i can see is the GPE Storm detected one. I waited 5-10 Minutes, nothing happens. Booting without acpi works - the system boots up. But.. as most things i use are using acpi, i cannot work with the system without acpi enabled. So i switched back to 2.6.25.8 for now. Any further information i could provide to help here? I'm not a kernel hacker, so no idea how i could help ;) Ah by the way, pressing the power button gives some information about a not connected battery - indeed the notebook battery isn't plugged in, don't know whether that's important or not. Created attachment 17102 [details]
/sys/firmware/acpi/interrupts $ grep . *
someone asked for this on a working kernel without the GPE storm detected message - so thats the output on a acer aspire 5520g using linux-2.6.25.8
On Wed, Aug 6, 2008 at 4:26 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #83 from jean.bruenn@ip-minds.de 2008-08-06 04:26 ------- > Created an attachment (id=17102) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17102&action=view) > /sys/firmware/acpi/interrupts $ grep . * > > someone asked for this on a working kernel without the GPE storm detected > message - so thats the output on a acer aspire 5520g using linux-2.6.25.8 > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > from reading the other posts, with the acer's they were having issues with function keys, typing was slow etc... nothing was pointing to battery issues(like the macbooks), but I could be wrong. One interesting post I had read was with the acer; using polling mode seems to work better than interrupt mode. So from just looking at the issue: the acers work better in polling mode, and the macbook's work better in interrupt mode. aside from this I've been using acpi_osi=Darwin for a while now, without having to touch anything with the kernel. doesn't help me; system isn't booting up with 2.6.26.1 using acpi_osi=Darwin (still the same problem then without) it's hanging after the GPE storm detected disabling .. line On Wed, Aug 6, 2008 at 11:21 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #85 from jean.bruenn@ip-minds.de 2008-08-06 11:21 ------- > doesn't help me; system isn't booting up with 2.6.26.1 using acpi_osi=Darwin > (still the same problem then without) it's hanging after the GPE storm > detected > disabling .. line > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Hmm.. The macbook pro I have is with the ati chipset. do you have the same? try taking the battery out so you can atleast boot up. (thats what I ended up having to do, as a short workaround) also make sure sbs.ko is either blacklisted, or not compiled, due to sbs hanging the system as well. i think you got me wrong, i have an acer aspire 5520 g, no macbook. completly nvidia chipset. On Wed, Aug 6, 2008 at 12:15 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #87 from jean.bruenn@ip-minds.de 2008-08-06 12:15 ------- > i think you got me wrong, i have an acer aspire 5520 g, no macbook. > > completly nvidia chipset. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I'm not sure how to handle an acer, with the gpe storm detector. i.g. kernel parameters, or hardware tweaks. you can try and revert the commit if need be so you can get up and running: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=d7a0e1f56472db0825e13f9dd39f0ad79b8c8b3e or change if (atomic_read(&ec->irq_count) > 5) { to if (atomic_read(&ec->irq_count) > 20) { in drivers/acpi/ec.c to prolong the storm from firing off. Hello, i tried the second thing - thats not working. I've seen a patch in BUG 11089 and tried it out - that fixes my problem (not booting) but brings up a new problem - Anyway, i can use that kernel now. ;-) On Wed, Aug 6, 2008 at 3:04 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #89 from jean.bruenn@ip-minds.de 2008-08-06 15:04 ------- > Hello, > > i tried the second thing - thats not working. I've seen a patch in BUG 11089 > and tried it out - that fixes my problem (not booting) but brings up a new > problem - Anyway, i can use that kernel now. ;-) > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Cool, glad you're up and running. (In reply to comment #89) > Hello, > > i tried the second thing - thats not working. I've seen a patch in BUG 11089 > and tried it out - that fixes my problem (not booting) but brings up a new > problem - Anyway, i can use that kernel now. ;-) > Hello, same problem on a vaio vgn-fe41s. if I boot with the AC adapter plugged the kernel freeze, when is freezed if I unplug the AC after a few seconds I read the GPE storm message. if I boot with the AC adapter unplugged everything works fine, also Fn and volume keys. I tried the same patch (BUG 11089) and now the situation is: if I boot with the AC adapter plugged the kernel starts normally, but the mute,vol up, vol down keys don't work, Fn keys still work. if I boot with the AC adapter unplugged the kernel freeze again, if I try to plug the AC after a few seconds I see the GPE storm message again On Mon, Aug 11, 2008 at 6:12 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > cafassomatte@libero.it changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |cafassomatte@libero.it > > > > > ------- Comment #91 from cafassomatte@libero.it 2008-08-11 06:12 ------- > (In reply to comment #89) >> Hello, >> >> i tried the second thing - thats not working. I've seen a patch in BUG 11089 >> and tried it out - that fixes my problem (not booting) but brings up a new >> problem - Anyway, i can use that kernel now. ;-) >> > > Hello, > same problem on a vaio vgn-fe41s. > > if I boot with the AC adapter plugged the kernel freeze, when is freezed if I > unplug the AC after a few seconds I read the GPE storm message. > > if I boot with the AC adapter unplugged everything works fine, also Fn and > volume keys. > > > I tried the same patch (BUG 11089) and now the situation is: > > if I boot with the AC adapter plugged the kernel starts normally, but the > mute,vol up, vol down keys don't work, Fn keys still work. > > if I boot with the AC adapter unplugged the kernel freeze again, if I try to > plug the AC after a few seconds I see the GPE storm message again > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Over here with a macbook if the GPE storm detector goes off, the system is still usable.(last I checked even battery info was still readable). As a work around acpi_osi=Darwin calms the interrupts down with the battery.(which seems to be the cause of the GPE storm, for some reason or another). The only problem, or complaint if you want to call it that, is when using acpi_osi=Darwin there is no battery info i.g. in a terminal typing "acpi" gives no reading. but I can run the system without having the GPE detector go off. As for using sbs.ko with the "Darwin" option, I seem to have a freeze, once the module is loaded. I was going to try and see if there was a patch to try or revert something, just to see if loading sbs.ko leads to something good. Alexey, do you have an update on the status of this bug? It doesn't look like any regression from 2.6.25 is still tracked here? [if there is an regression can someone please do a bisect to identify the commit that introduced it?]. It sounds more like this was always broken, correct? Also it seems like lots of people start posting unrelated issues in this bug. Folks, please open new bugs for new problems, even if they look similar on the first look. The bugs can be still merged then. On Sun, Aug 17, 2008 at 7:35 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > andi-bz@firstfloor.org changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |andi-bz@firstfloor.org > > > > > ------- Comment #93 from andi-bz@firstfloor.org 2008-08-17 19:35 ------- > Alexey, do you have an update on the status of this bug? > > It doesn't look like any regression from 2.6.25 is still tracked here? [if > there is an regression can someone please do a bisect to identify the commit > that introduced it?]. It sounds more like this was always broken, correct? > > Also it seems like lots of people start posting unrelated issues in this bug. > Folks, please open new bugs for new problems, even if they look similar on > the first look. The bugs can be still merged then. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Personally I kind of wish the gpe storm detector was commited earlier than it was, this way the issue would have been found earlier.(hmm.. maybe this explains why my battery crapped out after a year).on the macbook that is. Anyways I'm open to any ideas. Created attachment 17355 [details]
make poll switch manual
Don't do degrade to poll mode automatically, as per Pavels' suggestion.
On Thu, Aug 21, 2008 at 9:03 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #95 from astarikovskiy@suse.de 2008-08-21 09:03 ------- > Created an attachment (id=17355) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17355&action=view) > make poll switch manual > > Don't do degrade to poll mode automatically, as per Pavels' suggestion. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I'll try t out. On Thu, Aug 21, 2008 at 9:17 AM, Justin Mattock <justinmattock@gmail.com> wrote: > On Thu, Aug 21, 2008 at 9:03 AM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> >> >> >> ------- Comment #95 from astarikovskiy@suse.de 2008-08-21 09:03 ------- >> Created an attachment (id=17355) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=17355&action=view) >> make poll switch manual >> >> Don't do degrade to poll mode automatically, as per Pavels' suggestion. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > I'll try t out. > > -- > Justin P. Mattock > O.k. I've plugged in the patch attached are the different various boot params that I had used. As for any issues I'm seeing two issues 1)something with irq 23 nobody cared, causing me to be unable to login (typing a letter results in the letter being written twice i.g. pressing "z" will result in "zz". (this happens every 10 or so reboots); this is the original post that I had sent a few weeks ago: http://groups.google.com/group/fa.linux.kernel/browse_thread/thread/44deab2bc1ffeac3 2) loading uvcvideo without ift-load i.g. a few days ago I was receiving an error message with ift-load:(havent upgraded ift-load yet, so the issue might already be fixed). usb 5-1: usbfs: USBDEVFS_CONTROL failed cmd ift-load rqt 64 rq 160 len 1 ret -110 (attached is that just for a record) Anyways On a cold boot uvcvideo seems to not find isight.fw, until I start up the system, rmmod uvcvideo and then modprobe uvcvideo, then reboot. then everything is good. Now after plugging in this patch, for some reason I'm having the system freeze, after removing the module and loading again and rebooting.(strange); Overall after applying this patch it seems isight_firmware and uvcvideo need ift-load to work properly.(I just need to do an upgrade with ift-load). Sorry should of sent the attachments to the Bugreport, I'll do that right away justin P. Mattock On Aug 21, 2008, at 9:03 AM, bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #95 from astarikovskiy@suse.de 2008-08-21 09:03 > ------- > Created an attachment (id=17355) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17355&action=view) > make poll switch manual > > Don't do degrade to poll mode automatically, as per Pavels' > suggestion. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. Created attachment 17359 [details]
acpi_osi=Darwin+ec_intr=0
dmesg of applied don't degrade to poll mode patch
with acpi_osi=Darwin and ec_intr=0 as boot param
Created attachment 17360 [details]
acpi_osi=Darwin+ec_intr=non-0
dmesg of applied don't degrade to poll mode patch
with acpi_osi=Darwin and ec_intr=non-0 as boot param
Created attachment 17361 [details]
acpi_osi=Linux+ec_intr=0
dmesg of applied don't degrade to poll mode patch
with acpi_osi=Linux and ec_intr=0 as boot param
Created attachment 17362 [details]
acpi_osi=Linux+ec_intr=non-0
dmesg of applied don't degrade to poll mode patch
with acpi_osi=Linux and ec_intr=non-0 as boot param
Hmmm after looking at dmesg I'm noticing: ec_intr=non-0 is starting in poll mode if I don't put anything under my boot option I'll start in interrupt mode. I'm having some major issues also, system doesn't boot if acpi is on, passing acpi=off or disabling it in the menuconfig allows the system to boot. I've tried a number of 2.6.26 and 2.6.27 rc's, haven't found a fix yet. System is a HP dual core turion with an Nvidia chipset. George, Your issue is unrelated to this bug, so please either find more appropriate one or create new. Don't post here. Hi, Alexey It seems that there is no mode switch from interrput to polling mode after the patch in comment #95 hits the upstream kernel.Right? Is it reasonable? If on some laptops there is no GPE interrupt confirmation and it can't be switched to polling mode, in such case maybe EC can't work well. Maybe some laptops will be affected by this patch. Is it appropriate that the mode switch is disabled only for some specific laptops? Created attachment 17572 [details]
try the debug patch in which EC GPE works in level mode
Do someone have an opportunity to try the debug patch on the 2.6.27-rc4 kernel and see whether the EC GPE storm still exists?
Thanks.
On Tue, Sep 2, 2008 at 4:30 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #107 from yakui.zhao@intel.com 2008-09-02 01:30 ------- > Created an attachment (id=17572) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17572&action=view) > try the debug patch in which EC GPE works in level mode > > Do someone have an opportunity to try the debug patch on the 2.6.27-rc4 > kernel > and see whether the EC GPE storm still exists? > > Thanks. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I'll give it a try and see. I'll post the results as soon as I see any. regards; I've just tried the EDGE_TRIGGERED -> LEVEL_TRIGGERED patch on 2.6.27-rc5, on Fujitsu-Siemens Esprimo Mobile V5505 laptop. I still get "ACPI: EC: GPE storm detected, disabling EC GPE" on both x86_32 and x86_64: so far as I can tell this patch makes no difference. I've never noticed anything worse than the message itself, so my case is not a very important one. Created attachment 17587 [details]
try the debug patch in which the query_pending bit is cleared after processing EC notification event
Will you please try the debug patch on the latest kernel and see whether the number of ACPI interrupt is increased as fast as before?( cd /sys/firmware/acpi/interrupts/ ; grep . *)
In this debug patch the query_pending bit is cleared after EC notification event is processed. At the same time the source about detecting EC storm is disabled.
Thanks.
bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > yakui.zhao@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Attachment #17572 [details]|0 |1 > is obsolete| | > > > > > ------- Comment #110 from yakui.zhao@intel.com 2008-09-02 23:51 ------- > Created an attachment (id=17587) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17587&action=view) > try the debug patch in which the query_pending bit is cleared after > processing > EC notification event Tried that too... irq number is increased about 10x times. Regards, Alex. > > Will you please try the debug patch on the latest kernel and see whether the > number of ACPI interrupt is increased as fast as before?( cd > /sys/firmware/acpi/interrupts/ ; grep . *) > In this debug patch the query_pending bit is cleared after EC notification > event is processed. At the same time the source about detecting EC storm is > disabled. > > Thanks. > > bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #111 from astarikovskiy@suse.de 2008-09-02 23:57 ------- > bugme-daemon@bugzilla.kernel.org wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> yakui.zhao@intel.com changed: >> >> What |Removed |Added >> ---------------------------------------------------------------------------- >> Attachment #17572 [details]|0 |1 >> is obsolete| | >> >> >> >> >> ------- Comment #110 from yakui.zhao@intel.com 2008-09-02 23:51 ------- >> Created an attachment (id=17587) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17587&action=view) >> --> (http://bugzilla.kernel.org/attachment.cgi?id=17587&action=view) >> try the debug patch in which the query_pending bit is cleared after >> processing >> EC notification event > Tried that too... irq number is increased about 10x times. Oops. No, this approach does not change anything visible. There was a patch, where GPE is disabled while query is not services, it helped to reduce number of interrupts by about 3-5 times. >> Will you please try the debug patch on the latest kernel and see whether the >> number of ACPI interrupt is increased as fast as before?( cd >> /sys/firmware/acpi/interrupts/ ; grep . *) >> In this debug patch the query_pending bit is cleared after EC notification >> event is processed. At the same time the source about detecting EC storm is >> disabled. > >> Thanks. >> >> > > Hi, Alexey
Maybe this patch can't change anything visible. But the number of ACPI interrupt won't be increased as fast as before. On some laptops if the Query_pending bit is cleared too early, maybe one EC notification event is processed twice or more.
Thanks.
bugme-daemon@bugzilla.kernel.org wrote: > ------- Comment #113 from yakui.zhao@intel.com 2008-09-03 00:09 ------- > Hi, Alexey > Maybe this patch can't change anything visible. But the number of ACPI > interrupt won't be increased as fast as before. On some laptops if the > Query_pending bit is cleared too early, maybe one EC notification event is > processed twice or more. > Thanks. By "not visible" I mean also that number of interrupts does not change dramatically. And you are diverging from spec by doing this. > > There is no definition about how to process the EC notification event in ACPI spec. In the current upstream kernel when GPE interrupt is serviced, the GPE is deferred to be enabled again only after the GPE method is fully evaluated.(Including the notification event). Maybe it is also appropriate that the query_pending bit is cleared after EC notification event is processed.(_Qxx method is evaluated). In fact on my laptop the SCI_EVT bit is always set before the _Qxx method is evaluated. If the patch in comment #110 is applied, one EC notification event is processed three times. If the notification event is related with battery, maybe a lot of EC interrupts are at all meaningless. Created attachment 17609 [details]
Patch 1/2 : try the debug patch in which the query_pending bit is cleared after processing EC notification event
Created attachment 17611 [details]
Patch 2/2: try the debug patch in which EC work mode is simplified
Do someone have opportunity to try the attached two debug patches on the latest kernel(2.6.27-rc4/5) and see whether the system is stable?
In the patch 1: The query_pending bit will be cleared only after processing the EC notificatin event.
In the patch 2: The EC will work in polling mode when EC internal register is accessed. The EC GPE handler is only to process the EC notification event.
Thanks.
Created attachment 17617 [details]
Add some delay in EC GPE interrupt handler on some bios. (Patch 3)
Maybe the above two patches can't resolve the issue related with EC GPE storm.
Will you please try this patch together with the above two patches on the latest kernel and see whether the system is stable?
Thanks.
Ok, this one is really stupid. By adding delay to interrupt service routine, you halt the system for this time. Please stop if you don't understand what you are doing... Created attachment 17619 [details]
cleaner storm detection, lower lattencies
To keep you busy, here is rewrite of EC transaction mechanism, should allow faster transaction and lower latencies for queries. Also GPE is not disabled completely, so no queries are lost (even fast ones).
(In reply to comment #120) > Created an attachment (id=17619) [details] > cleaner storm detection, lower lattencies I tested the patch with 2.6.27-rc5 and didn't experience any GPE storm problems on the EeePC so far. On Mon, Sep 8, 2008 at 2:13 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #121 from florian-bug-kernel@heimgaertner.net 2008-09-08 > 14:13 ------- > (In reply to comment #120) >> Created an attachment (id=17619) > --> (http://bugzilla.kernel.org/attachment.cgi?id=17619&action=view) > [details] >> cleaner storm detection, lower lattencies > > I tested the patch with 2.6.27-rc5 and didn't experience any GPE storm > problems > on the EeePC so far. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > On a macbook or other? Hi, Florian
Do you have opportunity to try the patches in comment#116,117,118 on the kernel(2.6.27-rc5) and see whether the system can work well?
thanks.
On Mon, Sep 8, 2008 at 6:02 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #123 from yakui.zhao@intel.com 2008-09-08 18:02 ------- > Hi, Florian > Do you have opportunity to try the patches in comment#116,117,118 on the > kernel(2.6.27-rc5) and see whether the system can work well? > thanks. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.k. Ive tried all three of the patches attached is dmesg of the different boot params I used. overall I think I'm still seeing the same issue with the battery (lots of interrupts with acpi_osi=Linux option); When using this option the only unstable part I see is with unpluggin the A/C "takes a few seconds to dim, and become bright". noticed also a new avc's message from SELinux. Hi, Justin
Thanks for the test.
There also exist a lot of interrupts even when reading the battery info on most laptops.(Maybe about 100 EC GPE interrupts will be triggered). On your laptop the battery info is obtained by using SMBUS controller based on EC. Even when there is no EC GPE storm, about 1000 EC GPE interrupts will be triggered once reading the battery info.
At the same time several EC GPE interrupts will be triggered on your laptops although the EC pulse interrupt is triggered only once. In such case the situation will become more .
Does the system work well although lots of EC interrupts are triggered?
What does it mean that with unpluggin the A/C it will takes a few seconds to dim, and become bright?
From the acpidump it seems that the brightness is not controlled by ACPI video driver(there is no _BCL/_BQC_/_BCM object in DSDT table). Maybe it is controlled by BIOS. In such case whether it will take longer time to dim is not controlled by OS.
Thanks.
Hi, Justin
Thanks for the test.
There also exist a lot of interrupts even when reading the battery info on
most laptops.(Maybe about 100 EC GPE interrupts will be triggered). On your
laptop the battery info is obtained by using SMBUS controller based on EC. Even
when there is no EC GPE storm, about 1000 EC GPE interrupts will be triggered
once reading the battery info.
At the same time several EC GPE interrupts will be triggered on your laptops
although the EC pulse interrupt is triggered only once. In such case the
situation will become worse. Of course my workaround patch can reduce the number of EC GPE interrupts.
Does the system work well although lots of EC interrupts are triggered?
What does it mean that with unpluggin the A/C it will takes a few seconds
to dim, and become bright?
From the acpidump it seems that the brightness is not controlled by ACPI
video driver(there is no _BCL/_BQC_/_BCM object in DSDT table). Maybe it is
controlled by BIOS. In such case whether it will take longer time to dim is not
controlled by OS.
Thanks.
On Mon, Sep 8, 2008 at 8:21 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #125 from yakui.zhao@intel.com 2008-09-08 20:21 ------- > Hi, Justin > Thanks for the test. > There also exist a lot of interrupts even when reading the battery info on > most laptops.(Maybe about 100 EC GPE interrupts will be triggered). On your > laptop the battery info is obtained by using SMBUS controller based on EC. > Even > when there is no EC GPE storm, about 1000 EC GPE interrupts will be triggered > once reading the battery info. > At the same time several EC GPE interrupts will be triggered on your > laptops > although the EC pulse interrupt is triggered only once. In such case the > situation will become more . > > Does the system work well although lots of EC interrupts are triggered? > > What does it mean that with unpluggin the A/C it will takes a few seconds > to dim, and become bright? > From the acpidump it seems that the brightness is not controlled by ACPI > video driver(there is no _BCL/_BQC_/_BCM object in DSDT table). Maybe it is > controlled by BIOS. In such case whether it will take longer time to dim is > not > controlled by OS. > Thanks. > > > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Yes the system is usable in the case of a gpe storm etc.. as for the dimming part when using a macbook I use an app called pommed, when unplugging the A/C adapter to go on "battery", pommed will automatically dim the backlight. when the gpe storm is triggered this is when I will se a longer pause when dimming than normally. (thats all); in any case I don't think pommed is as big of an issue as for the other posts I was reading. In my case I think I need to look at the dsdt and see if I can find anything in there that might be of use. Hi, Justin
Is the system stable after applying the patches in comment #116,117,118?
Understand what you said. When unplugging the A/C adapter, maybe the user space application called pommed will check the status of the battery and change the brightness of backlight. In such case it will take more time to read the info of the battery. Can you confirm whether the battery info is obtained again after unplugging AC adapter?( by checking whether the number of EC GPE interrupts are increased very fast. cd /sys/firmware/acpi/interrupts; grep . *)
Maybe it will be very quick if the pommed application changes the brightness of backlight only according to the status of AC adapter.
Thanks.
Oh shit, I just did Git-reset --hard origin Git-pull I'll have to do that again, to look at the Interrupts. Although I did glance And saw the interrupt was at 1913 Didn't see how many per second or minute(I'll get back to you with that); Now, when looking at my other Linux box(dell inspiron 1200),even though Hardware is different, Im seeing 1 Interrupt coming from gpe1D and gpe1C should this be similar with A MacBook? Or is this something With sbs.ko sbshc.ko? That is loading With the dell box. justin P. Mattock On Sep 8, 2008, at 8:54 PM, bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #128 from yakui.zhao@intel.com 2008-09-08 20:54 > ------- > Hi, Justin > Is the system stable after applying the patches in comment > #116,117,118? > Understand what you said. When unplugging the A/C adapter, maybe > the user > space application called pommed will check the status of the battery > and change > the brightness of backlight. In such case it will take more time to > read the > info of the battery. Can you confirm whether the battery info is > obtained again > after unplugging AC adapter?( by checking whether the number of EC GPE > interrupts are increased very fast. cd /sys/firmware/acpi/ > interrupts; grep . *) > Maybe it will be very quick if the pommed application changes the > brightness of backlight only according to the status of AC adapter. > Thanks. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. This is related with sbs.ko. In fact on the macbootk laptop it is unnecessary to load SBS driver again(sbs.ko, sbshc.ko). From the acpidump it seems that the battery info is obtained by using the SMbus controller(this is realized in AML code). If the sbs driver is loaded, maybe the sbs driver will register another battery device, which is the same as the device registered by battery driver.
Please don't load the sbs driver on the macbook laptop.
On Mon, Sep 8, 2008 at 10:34 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #130 from yakui.zhao@intel.com 2008-09-08 22:34 ------- > This is related with sbs.ko. In fact on the macbootk laptop it is > unnecessary to load SBS driver again(sbs.ko, sbshc.ko). From the acpidump it > seems that the battery info is obtained by using the SMbus controller(this is > realized in AML code). If the sbs driver is loaded, maybe the sbs driver will > register another battery device, which is the same as the device registered > by > battery driver. > > Please don't load the sbs driver on the macbook laptop. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Thanks for the info on sbs.ko,(no wonder it has a hard time loading); Anyways I went back and plugged you're patch in after updating git., attached you will find: /sys/firmware/acpi/interrupts/grep . * four types with the different boot params etc.. From looking at the info the interrupt number could be high when using the patch due to the gpe storm going off, as opposed to the gpe storm not going off in that mode(at least thats what I can see from here); As for watching the interrupt, using: watch /sys/firmware/acpi/interrupts/grep . * gave me a nice perspective, when under acpi_osi=Linux option the number jumps 2X's by 100 +- and nothing for around 30 +- seconds, and then jumps up after a short break. I should try and post you the results of grep .* when the gpe storm is detected and when it isn't Justin, Could you please also test the patch from comment #120? On a NEC Versa M370, 116+117 did not help, with 120 it looks fine. I did not try 118 yet. On Tue, Sep 9, 2008 at 3:14 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #133 from pterjan@gmail.com 2008-09-09 03:14 ------- > On a NEC Versa M370, 116+117 did not help, with 120 it looks fine. I did not > try 118 yet. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Yeah I can give it a try. Ill undo the three patches and then try #120 then after that I can try #118 (In reply to comment #133) > On a NEC Versa M370, 116+117 did not help, with 120 it looks fine. I did not > try 118 yet. thanks for the test. It seems that the system with the patch in comment #120 can work well. It seems that there also exists the issue of EC GPE interrupt storm on your system. Will you please try the patches in comment#116,117,118? on the latest kernel of 2.6.27-rc5?( If the patch in comment#118 is not used, it can't workaround the EC GPE storm). Thanks. Sorry for being so long, it seems to be fixed with either 116+117+118 or 120 on 2.6.27-rc5-git7 on the NEC Versa M370. On Tue, Sep 9, 2008 at 9:01 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #136 from pterjan@gmail.com 2008-09-09 09:01 ------- > Sorry for being so long, it seems to be fixed with either 116+117+118 or 120 > on > 2.6.27-rc5-git7 on the NEC Versa M370. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > yesterday I used 116+117+118 on 2.6.27-rc5 now I'll plug in 120 and post results of /sys/firmware/acpi/interrupts/grep . * and dmesg. is there another place in the kernel where I should look and possibly record the data, so It help's you guys out? On Tue, Sep 9, 2008 at 9:46 AM, Justin Mattock <justinmattock@gmail.com> wrote: > On Tue, Sep 9, 2008 at 9:01 AM, <bugme-daemon@bugzilla.kernel.org> wrote: >> http://bugzilla.kernel.org/show_bug.cgi?id=10724 >> >> >> >> >> >> ------- Comment #136 from pterjan@gmail.com 2008-09-09 09:01 ------- >> Sorry for being so long, it seems to be fixed with either 116+117+118 or 120 >> on >> 2.6.27-rc5-git7 on the NEC Versa M370. >> >> >> -- >> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email >> ------- You are receiving this mail because: ------- >> You are on the CC list for the bug, or are watching someone who is. >> > > yesterday I used 116+117+118 on 2.6.27-rc5 > now I'll plug in 120 and post results of /sys/firmware/acpi/interrupts/grep . > * > and dmesg. is there another place in the kernel where I should look and > possibly record the data, so It help's you guys out? > > -- > Justin P. Mattock > O.k. took me a bit longer to plug this patch in #120(forgot to put "int" in a line); but I manage to do so. attahced is dmesg and, /sys/firmware/acpi/interrupts/grep . * Alexey was so friendly to send me ec.c patched with the patch from #120. Thanks a lot. I have tested it on Asus EeePC model 701. The reaction on the Eee's special keys got much quicker and a *lot* more stable. The system is very usable with this version now. (It was very unstable before.) If I insist *really* hard I am still able to block reaction to these special keys by repeatedly sending a lot of these key events (by autorepeat). I cannot find any error messages after that event. It looks to me as if the hardware might do that on his own and the kernel might not even see it. Is there anything I can do to debug the situation? Regarding our second problem with ec.c, the bogus thermal shutdowns during boot process (showing on some EeePC models) the version you sent me seems rock stable. This is probably due to the change introduced by commit http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9d699ed92a459cb408e2577e8bbeabc8ec3989e1 Robert Epprecht Additional info regarding comment #139 | Regarding our second problem with ec.c, the bogus thermal shutdowns during boot | process (showing on some EeePC models) the version you sent me seems rock | stable. This is probably due to the change introduced by commit | http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9d699ed92a459cb408e2577e8bbeabc8ec3989e1 Sorry, forgot to mention that i had found this using git bisect. The commit does not talk about thermal shutdown problems, but it *fixes* them (on the Asus Eee). Using git bisect I also found the point where this problems with bogus thermal shutdowns of Asus EeePCs started. First broken version was after the following change http://git.kernel.org/?p=linux/kernel/git/ak/linux-acpi-2.6.git;a=commitdiff;h=1b7fc5aae8867046f8d3d45808309d5b7f2e036a Robert Epprecht On Fri, Sep 19, 2008 at 2:53 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #140 from epprecht@solnet.ch 2008-09-19 02:53 ------- > Additional info regarding comment #139 > > | Regarding our second problem with ec.c, the bogus thermal shutdowns during > boot > | process (showing on some EeePC models) the version you sent me seems rock > | stable. This is probably due to the change introduced by commit > | > > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9d699ed92a459cb408e2577e8bbeabc8ec3989e1 > > Sorry, forgot to mention that i had found this using git bisect. > The commit does not talk about thermal shutdown problems, but it *fixes* them > (on the Asus Eee). > > Using git bisect I also found the point where this problems with bogus > thermal > shutdowns of Asus EeePCs started. First broken version was after the > following > change > > http://git.kernel.org/?p=linux/kernel/git/ak/linux-acpi-2.6.git;a=commitdiff;h=1b7fc5aae8867046f8d3d45808309d5b7f2e036a > > Robert Epprecht > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Glad to see everything is working for you. Overall I think this is an intricate issue due to different types of machines. with using a macbook there was no keyboard issues, or thermal shutdowns (at least from what I can see), just something with the battery. Created attachment 18044 [details] patch vs 2.6.27-rc7 this version of Alexey's patch of comment #120 has been checked into the acpi test tree. If it doesn't work for you, please speak up. *** Bug 11679 has been marked as a duplicate of this bug. *** On Wed, Oct 1, 2008 at 1:10 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #143 from rjw@sisk.pl 2008-10-01 13:10 ------- > *** Bug 11679 has been marked as a duplicate of this bug. *** > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I was going to say something, but then was carried away with another task, and totally forgot. > I was going to say something, but then was carried away
> with another task, and totally forgot.
does the patch help, or does it not help?
On Thu, 2008-10-16 at 22:23 -0700, bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #145 from len.brown@intel.com 2008-10-16 22:23 ------- > > I was going to say something, but then was carried away > > with another task, and totally forgot. > > does the patch help, or does it not help? > > O.K. I applied the patch for 2.6.27 attached is dmesg of: withpatch and withoutpatch applied > attached is dmesg of: withpatch and without patch applied attached to what? e-mail to bugzilla drops attachments. Please go here: http://bugzilla.kernel.org/attachment.cgi?bugid=10724&action=enter and upload the files. But don't keep me in suspense, does the patch help, or does it not help? Alright; I'll upload the attachments later(in the process of writting an SELinux policy) as for the system itself I think you nailed it, usually I Would see the gps storm fire of within Minutes this time nothing, as for any breakage the touchpad is non responsive(not sure if it was from the patch or an update through apt); I am seeing something with the Atheros module, not sure if related. But overall I think this does it. As soon as I finish with the policy I'll let the system sit running to see if anything happens. Let me know what info you need and I'll Try and post it to you as soon as possible, as for applying the patch There was a / missing in a comment And a + to a thread, but it was late when I plugged it in could of been me justin P. Mattock On Oct 17, 2008, at 11:05 AM, bugme-daemon@bugzilla.kernel.org wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #147 from len.brown@intel.com 2008-10-17 11:05 > ------- >> attached is dmesg of: withpatch and without patch applied > > attached to what? e-mail to bugzilla drops attachments. > Please go here: > http://bugzilla.kernel.org/attachment.cgi?bugid=10724&action=enter > and upload the files. > > But don't keep me in suspense, > does the patch help, or does it not help? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. On Fri, Oct 17, 2008 at 2:30 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > len.brown@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Component|Platform-Drivers |EC > > > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > after running the system all day and recompiling refpolicy atleast 100 times, I still am not receiving a gpe storm. there was a moment in of which I thought it was triggered, but no. this was what I noticed with the patch: * OpReg are installed */ after looking at the patch this was the only problem I can see (the other issue was my imagination); I'll try and post dmesg ASAP for the record. Created attachment 18358 [details]
dmesg of 2.6.27-rc7 patch applied to a macbook pro
This patch seems pretty good, what are the results of
other laptop manufacturers
this fix break the fix of bug 9823 (http://bugzilla.kernel.org/show_bug.cgi?id=9823) on MSI PR200/System76 daru2 these notebooks needs the ec poll mode but it's started in interrupt mode. is there a way to enable poll mode via boot option? On Sat, Oct 18, 2008 at 3:19 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > timo.tretter@student.hpi.uni-potsdam.de changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |timo.tretter@student.hpi.uni > | |-potsdam.de > > > > > ------- Comment #151 from timo.tretter@student.hpi.uni-potsdam.de 2008-10-18 > 03:19 ------- > this fix break the fix of bug 9823 > (http://bugzilla.kernel.org/show_bug.cgi?id=9823) on MSI PR200/System76 daru2 > > these notebooks needs the ec poll mode but it's started in interrupt mode. > is there a way to enable poll mode via boot option? > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > I can't remember which one, but when I applied one of the patches before the 2.6.27 patch, my system was booting into polling mode, but this was the effects of a mac, not sure how other hardware reacts. As for a boot param I think there is one, or it might of been one of the patches i.g. "ec_intr=1" (In reply to comment #151) > this fix break the fix of bug 9823 > (http://bugzilla.kernel.org/show_bug.cgi?id=9823) on MSI PR200/System76 daru2 If this is the case, please re-open bug #9823. Please provide acpidump and dmesg outputs with DEBUG enabled in drivers/acpi/ec.c. On Sat, Oct 18, 2008 at 7:26 AM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > > > > ------- Comment #153 from astarikovskiy@suse.de 2008-10-18 07:26 ------- > (In reply to comment #151) >> this fix break the fix of bug 9823 >> (http://bugzilla.kernel.org/show_bug.cgi?id=9823) on MSI PR200/System76 >> daru2 > If this is the case, please re-open bug #9823. Please provide acpidump and > dmesg > outputs with DEBUG enabled in drivers/acpi/ec.c. > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > Im sorry my bad; there nothing wrong with the 2.6.27 patch(* OpReg are installed */) I needed to look up a line.... and take a break.... shipped in linux-2.6.28-rc1 closed commit 7c6db4e050601f359081fde418ca6dc4fc2d0011 Author: Alexey Starikovskiy <astarikovskiy@suse.de> Date: Thu Sep 25 21:00:31 2008 +0400 ACPI: EC: do transaction from interrupt context On Fri, Oct 24, 2008 at 11:23 PM, <bugme-daemon@bugzilla.kernel.org> wrote: > http://bugzilla.kernel.org/show_bug.cgi?id=10724 > > > len.brown@intel.com changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Status|RESOLVED |CLOSED > > > > > ------- Comment #155 from len.brown@intel.com 2008-10-24 23:23 ------- > shipped in linux-2.6.28-rc1 > closed > > commit 7c6db4e050601f359081fde418ca6dc4fc2d0011 > Author: Alexey Starikovskiy <astarikovskiy@suse.de> > Date: Thu Sep 25 21:00:31 2008 +0400 > > ACPI: EC: do transaction from interrupt context > > > -- > Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > O.k. reverted the patch, and pulled the latest git I'll let you know if I see anything. |