Bug 5161
Summary: | Regression: (2.6.12.5 -> 2.6.13) ACPI-0700 Error acpi_ev_gpe_dispatch - HP Compaq NX 6110 | ||
---|---|---|---|
Product: | Drivers | Reporter: | Olivier Mondoloni (elgrande71) |
Component: | PCI | Assignee: | Shaohua (shaohua.li) |
Status: | CLOSED PATCH_ALREADY_AVAILABLE | ||
Severity: | blocking | CC: | acpi-bugzilla, torvalds, yaq |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.13 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
acpidump
dmidecode proc interrupts lspci -vv screen copy of error messages dmesg from a working kernel 2.6.12 with acpi debug activated dmesg from 2.6.13 with acpi debug screen copy after patching 2.6.13 dmesg from 2.6.13 rc1 dmesg from 2.6.13 rc2 dmesg from 2.6.13 rc3 dmesg from 2.6.13 rc4 dmesg from 2.6.13 rc5 dmesg from 2.6.13 rc6 dmesg from 2.6.13 rc7 Screen foto before dump_stack() Dump_stack output. dmesg from 2.6.13 patch error messages while patching dmesg from booting 2.6.13 patched |
Description
Olivier Mondoloni
2005-08-31 06:50:29 UTC
Created attachment 5825 [details]
acpidump
acpidump hp nx 6110
Created attachment 5826 [details]
dmidecode
dmidecode hp nx 6110
Created attachment 5827 [details]
proc interrupts
proc interrupts hp nx 6110
Created attachment 5828 [details]
lspci -vv
lspci -vv hp nx 6110
Created attachment 5829 [details]
screen copy of error messages
What's my computer displays.
ec_burst=1 or ec_polling boot parameters don't fix the problem. ec_burst=1 don't fix the problem. same thing for ec_polling. I have the same problem: hp nx6110, gentoo, kernel 2.6.13. When setting in kernel .config: CONFIG_ACPI=y CONFIG_ACPI_BOOT=y CONFIG_ACPI_INTERPRETER=y # CONFIG_ACPI_SLEEP is not set # CONFIG_ACPI_AC is not set # CONFIG_ACPI_BATTERY is not set # CONFIG_ACPI_BUTTON is not set # CONFIG_ACPI_VIDEO is not set # CONFIG_ACPI_HOTKEY is not set # CONFIG_ACPI_FAN is not set # CONFIG_ACPI_PROCESSOR is not set # CONFIG_ACPI_ASUS is not set # CONFIG_ACPI_IBM is not set # CONFIG_ACPI_TOSHIBA is not set CONFIG_ACPI_BLACKLIST_YEAR=0 # CONFIG_ACPI_DEBUG is not set CONFIG_ACPI_BUS=y CONFIG_ACPI_EC=y CONFIG_ACPI_POWER=y CONFIG_ACPI_PCI=y CONFIG_ACPI_SYSTEM=y # CONFIG_X86_PM_TIMER is not set # CONFIG_ACPI_CONTAINER is not set it is possible to boot, but it hangs on halt or reboot. Everything works fine on 2.6.12.5. In /proc/ioports, you will find an entry of 'GPE0_BLK', see the number ahead is x. Could you please tell me the output of 'inl x' and 'inl x+4'? Please do it in the unworking case. Is there any chance we can get the full boot log from the system by a serial port? cat /proc/ioports gives me 1028-102f GPE0_BLK. inl x gives 0xFFFFFFFF inl x+4 gives 0xFFFFFFFF. Sorry but it's not possible for me to give you full boot log via serial port. Created attachment 5839 [details]
dmesg from a working kernel 2.6.12 with acpi debug activated
Look at the following line in this boot log file :
evgpeblk-0308 [08] ev_save_method_info ...
Is it related to my problem ?
Created attachment 5840 [details]
dmesg from 2.6.13 with acpi debug
several error messages repeated.
>evgpeblk-0308 [08] ev_save_method_info ...Is it related to my problem ?
This one doesn't matter, it's quite common in HP latops.
inl 0x1028 gives 0xFFFFFFFF and inl 0x102c gives 0xFFFFFFFF, right? It's quite
strange, all GPEs are enabled, which causes the error messages.
The output should be different in 2.6.12, right? (just want to confirm we
looked at the correct registers.)
Could we know where is the first error message? Eg, you might just disable the
error message after one shot
you can add:
--- evgpe.c 2005-08-29 10:18:00.000000000 +0800
+++ evgpe.c.new 2005-09-01 14:37:12.926836648 +0800
@@ -698,6 +698,10 @@
ACPI_REPORT_ERROR ((
"acpi_ev_gpe_dispatch: No handler or method for GPE[%
2X], disabling event\n",
gpe_number));
+ {
+ dump_stack();
+ while(1);
+ }
/*
Created attachment 5844 [details]
screen copy after patching 2.6.13
After patching my kernel, my computer startup was stopped by kernel crash (look
at the following attachement : ecran3acpierr.jpg).
I want to add that my kernel 2.6.13 was patched before with the lastest acpi
patch (acpi-20050815-2.6.13.diff.bz2).
Thanks.
Looks a RTC interrupt is generated, but the system thinks it's a GPE wrongly. But I have no idea what's wrong here. Could you please narrow down the issue (see which version of kernel start has the issue you might start from 13-rc1)? Thanks! Created attachment 5860 [details]
dmesg from 2.6.13 rc1
no problems with it.
Created attachment 5861 [details]
dmesg from 2.6.13 rc2
no problems with it.
Created attachment 5862 [details]
dmesg from 2.6.13 rc3
no problems with it.
Created attachment 5863 [details]
dmesg from 2.6.13 rc4
no problems with it.
Created attachment 5864 [details]
dmesg from 2.6.13 rc5
no problems with it.
Created attachment 5865 [details]
dmesg from 2.6.13 rc6
no problems with it.
Created attachment 5866 [details]
dmesg from 2.6.13 rc7
no problems with it.
I have no problems with 2.6.13 rc1 to 2.6.13 rc7 included. The problems come from 2.6.13 only. Thanks for help. Created attachment 5868 [details]
Screen foto before dump_stack()
nx6110 has no serial, USB serial console starts after acpi so this is best i
could do. Kernel 2.6.13, no patches
Created attachment 5869 [details]
Dump_stack output.
This was separate boot. Output is not always the same, next time I've tried to
boot, it didn't fit in the sreen.
2.6.13-rc7 works fine.
I have the same problem on a HP nw8240.. Anything I can do to help? Hi everybody, Many thanks for your test! It seems some yenta devices now wrongly use the memory of ACPI. We previously reserve ioports of motherboard and then allocate resources to PCI devices. But current PCI code seems violate the rule, which causes motherboard resources allocates to PCI devices. (The ACPI should use ioport 0x1000 - 0x107f, but in Maciej's log, the ioport from 0x1000-0x10ff now is allocated to cardbus bridge) Adding Linus to the thread, he generally has many ideas about such issues. >It seems some yenta devices now wrongly use the memory of ACPI.
Sorry, it should be ioport.
Looks it has been fixed by this patch: http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux- 2.6.git;a=commitdiff_plain;h=81d4af1340badcd2100c84fbd1bfd13156de41aa;hp=ae11be 6f37dfa140b6c2d1a53b307ef57da69a04 can you confirm it? Created attachment 5900 [details]
dmesg from 2.6.13 patch
It works great with the patch previously applied.
Thank you for your great job.
Created attachment 5901 [details]
error messages while patching
I have been obliged to apply the patch manually because of errors during the
patching process.
Look at the attachement.
I have the same problem with an HP Compaq nc4200 and I tried the patch : http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff_plain;h=81d4af1340badcd2100c84fbd1bfd13156de41aa;hp=ae11be6f37dfa140b6c2d1a53b307ef57da69a04 and now my computer hangs at boot time in another way : PCI: Cannot allocate region 7 of bridge 0000:00:1c.1 PCI: Cannot allocate region 8 of bridge 0000:00:1c.1 PCI: Cannot allocate region 9 of bridge 0000:00:1c.1 Created attachment 5903 [details]
dmesg from booting 2.6.13 patched
2.6.13 patched with above patch, boots fine, no errors during boot or patching.
Philippe, please report your issue to PCI guys. Anyway, this is a PCI bug and we have fix already. Closing. |