Hi,
在 2025/01/27 20:45, Greg KH 写道:
On Mon, Jan 27, 2025 at 10:11:30AM +0100, Greg KH wrote:
On Mon, Jan 27, 2025 at 04:49:23PM +0800, Yu Kuai wrote:
This set fix reported problem:
https://lore.kernel.org/all/CAJpMwyjmHQLvm6zg1cmQErttNNQPDAAXPKM3xgTjMhbfts986Q@xxxxxxxxxxxxxx/
https://lore.kernel.org/all/ADF7D720-5764-4AF3-B68E-1845988737AA@xxxxxxxxxxxxxxx/
See details in patch 5.
Why were these all not properly taggeed for stable inclusion to start
with?
Because the orignal problem in raid5 is not 100% clear, I don't know
what's the exact fix tag here, raid5 is quite complicated :(
So this set fix the problem by refactor, removing all related code in
raid5, and reimplement the logical in md.c. Please notice this refactor
can also improve performance for raid5.
I can't reporduce this problem myself, however, it's tested by other
folks that v4.19 is still affected, while I don't think this *refactor*
will be possible in that version. I'll suggest people to upgrade their
kernel to v6.6+ first if they're really affected...
Also, as these are not in a released kernel just yet, why should we
include them in one now before 6.14-rc1 is out?
I just want to send v6.6 version before the Chinese New Year, becaue
6.14-rc1 will be released while I'm on vacation.
Thanks,
Kuai
thanks,
greg k-h
.
