Bug 1533
Summary: | Unalignment access | ||
---|---|---|---|
Product: | ACPI | Reporter: | Luming Yu (luming.yu) |
Component: | ACPICA-Core | Assignee: | Robert Moore (Robert.Moore) |
Status: | CLOSED CODE_FIX | ||
Severity: | normal | CC: | acpi-bugzilla |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | Lastest 2.4 ia64 + Latest ACPI | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
a patch for fixing this issue
a patch for fixing this issue unaligned access call trace |
Description
Luming Yu
2003-11-12 21:32:03 UTC
Created attachment 1475 [details]
a patch for fixing this issue
acglobal.h is platform independent, but your method is platform dependent. So I guess you should find another solution. I'd like to see a call trace of this problem. The acpi_generic_address is compiled under #pragma pack(1) already, and we carefully extract data from it. This is not to say that a case has been missed -- and the call trace would help with that. But gcc generate unaligned address like below: [root@acpi-tiger 2.4-acpi]# grep acpi_gbl_xpm1a System.map e000000004cbde89 B acpi_gbl_xpm1a_enable Created attachment 1711 [details]
a patch for fixing this issue
This patch is based on Bob's suggestion
Fix integrated into 20040220 2.6 kernel still has this issue on IPF. So reopen this. unaligned access happens both on boot and sci triggered. pls, see the call trace attached. Created attachment 3680 [details]
unaligned access call trace
The call trace points toward something to do with the GPEs, but there is not enough info to go further than that. You will need to debug it down to the exact problem. Bob fixed this in ACPICA 20041015. Should be fixed ACPICA 20041015 shipped in 2.6.10-rc2 - closing. |