Bug 1653 - _GTS must be executed just prior to setting the sleep enable bit in PM1
Summary: _GTS must be executed just prior to setting the sleep enable bit in PM1
Status: REJECTED WILL_NOT_FIX
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Sleep-Wake (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Robert Moore
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-12-08 00:07 UTC by Luming Yu
Modified: 2005-09-29 14:13 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6
Subsystem:
Regression: ---
Bisected commit-id:

Attachments
a patch for fixing this issue (1.32 KB, patch)
2003-12-08 00:09 UTC, Luming Yu 		Details | Diff
Add an attachment (proposed patch, testcase, etc.)

Description Luming Yu 2003-12-08 00:07:56 UTC 
ACPI spec 2.0b section 7.3.3 \GTS said:
OSPM will set the sleep enable (SLP_EN) bit in PM1 control register immediately
following the execution of _GTS control method without performing any other I/O
or  allowing any interrupt servicing.

Currently, kernel do like below:

  _GTS ...
  device_suspend(state)
  ..
  acpi_enter_sleep_state(state)
Comment 1 Luming Yu 2003-12-08 00:09:41 UTC 
Created attachment 1630 [details]
a patch for fixing this issue

just move _GTS to right place.
Comment 2 Robert Moore 2004-07-23 10:53:20 UTC 
This may require more investigation.  Apparently "Other" ACPI implementations 
do not do this, and they may never do this, regardless of what the spec says.
Bob
Comment 3 Len Brown 2004-11-15 21:48:49 UTC 
perhaps we should make calling _GTS configurable then, 
in case it works on some BIOS and doesn't on others?
Comment 4 Robert Moore 2005-09-29 14:13:28 UTC 
Unless this proves to be a problem on a particular machine, I would rather not 
change the code, especially since we believe that it currently behaves exactly 
the same as Windows.
Note You need to log in before you can comment on or make changes to this bug.