Skip to contentSkip to conversation

libata-dev.git rebased; ACPI turned on

DISCUSSION Jeff Garzik 1 participant 1 message Oct 5, 2007 libata

lore
Share & tools
lore
#0 lore raw

1) I just rebased libata-dev.git and all its branches. If you are unfamiliar with git rebasing, this means you must either re-clone, or force-update your branch like this:

# grab latest nv-swncq branch
URL=git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev.git
cd /local/repo/libata-dev
git-checkout master	# or any branch other than what
			# is being updated
git-fetch -f $URL nv-swncq:nv-swncq	# force update of 'nv-swncq'

2) libata ACPI support has been turned on by default. So let's be alert for problems related to that. Problems will _probably_ fall into one of those categories:
* suspend/resume glitches
* problems related to the new ATA commands we are sending to the drive, as passed to us from the BIOS ACPI tables

On the upside, this should make libata suspend/resume a lot more sane.

3) I had to yank the FIFO drain patch. Alan says that the FIFO drain patch that went in causes problems on some chips, and that we cannot apply it unconditionally to all SFF controllers. He recommends a new ->drain_fifo() hook, with a default ata_sff_drain_fifo() implementation matching what Mark Lord previously submitted.

Jeff

P.S. I still have some more patches to apply, but they should not require a rebase.