[Pkg-sysvinit-devel] Please test the libata-fixes branch
kel at otaku42.de
Wed Jun 27 11:35:46 UTC 2007
The patch from r1054 that did global inhibition of -h halt option when
`ls -1 /sys/class/scsi_disk/*/manage_start_stop' was true seems like an easy
way out, but not a foolproof solution. That may not be true in all cases: one
disk may have libata management of write flush and spindown, while another
disk in the same system could conceivably not have that support.
Tejun Heo hinted that a suse developer, Werner Fink, was working on a
more substantial effort to update hddown.c logic, providing per disk device
decision on whether or not to give standby signal and cause flush of
write-cache or not based on sysfs detection of attached disk devices.
Attached is diff against the libata branch of pkg-sysvinit that adds Werner's
last patch. I thought it was at least worth a go on my own hardware, if
only for my own curiosity's sake. Seems to do what it advertises too.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 14846 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/pkg-sysvinit-devel/attachments/20070627/7f462b6c/attachment-0001.patch
More information about the Pkg-sysvinit-devel