Skip to content

PM_EVENT_PRETHAW

David Brownell linux-pm 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest6 patches 10 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/6fix broken/dubious driver suspend() methodsdrivers/ide/ppc/pmac.c drivers/media/dvb/cinergyT2/cinergyT2.c drivers/scsi/libata-core.c +1 more+16 -131 reply
2/6add PM_EVENT_PRETHAWinclude/linux/pm.h+48 -162 replies
3/6PM_EVENT_PRETHAW, handle in IDE and PCIdrivers/ide/ide.c drivers/pci/pci.c+7 -32 replies
4/6PM_EVENT_PRETHAW for various graphics cardsdrivers/video/aty/radeon_pm.c drivers/video/i810/i810_main.c drivers/video/nvidia/nvidia.c +1 more+30 -244 replies
5/6PM_EVENT_PRETHAW, handle for USBdrivers/usb/core/hcd-pci.c drivers/usb/core/usb.c drivers/usb/host/ehci-pci.c +3 more+24 -40 replies
6/6PM_EVENT_PRETHAW, issue from PM corekernel/power/disk.c kernel/power/swsusp.c kernel/power/user.c+11 -41 reply

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [patch/rft 2.6.17-rc5-git 0/6] PM_EVENT_PRETHAW

Following this message will be patches adding the new PRETHAW event, so that drivers can make sure that swsusp doesn't put hardware into bogus states before resuming.

As previously discussed, this is needed by drivers which examine the hardware state during resume() methods ... notably, USB controller drivers, which expect to see true suspend states in order to handle remote wakeup, but currently break with swsusp. Only two states are valid in resume(): after hardware reset, or else the state that suspend() left it in. PRETHAW allows drivers to force the former.

In sequence, the patches are:

- prethaw-misc.patch ... fixes code that's currently broken/dubious so that it won't care about the new message (and is thus more or less mergeable regardless of the rest of these patches)

- prethaw-header.patch ... defines the new event and its message

- prethaw-fw.patch ... IDE and (dumb) PCI can handle it simplistically

- prethaw-video.patch ... likewise, but this is per-driver not at the framework level

- prethaw-usb.patch ... fixing various "swsusp resume fails if the HCD is statically linked" bugs

- prethaw-core.patch ... updating the pm core to issue PRETHAW events, handling for which which the previous patches added.

Yes, it might be worth splitting some of those driver patches out into patch-per-driver format. Yes, I _did_ look at a couple hundred drivers (grr) to see what needed changing ... darn few drivers treat suspend() as anything other than PM_EVENT_SUSPEND.

- Dave