r268075 (MFV r267565) undone FreeBSD-specific changes

Andriy Gapon avg at FreeBSD.org
Wed Sep 17 10:53:06 UTC 2014


On 17/09/2014 11:52, Andriy Gapon wrote:
> So, it seems that r256889 is definitely undone now.  I have not checked if there
> is any other fallout from removing minblocksize parameter of
> zio_compress_data().  Hmm, in fact, it looks like zio_compress_data() in
> zio_write_bp_init() has suffered the same fate: whereas previously minblocksize
> was set metaslab_class_get_minblocksize(mc), now SPA_MINBLOCKSIZE is used for
> rounding up.

After more analysis I think that the changes in zio_write_bp_init() are correct.
 psize needs to be only SPA_MINBLOCKSIZE aligned.  The code in
zio_vdev_io_start() will take care of adjusting for ashift.

On the other hand, the change in l2arc_compress_buf() is definitely incorrect.
l2arc code uses physical ZFS I/O and as such it must correctly set data size and
data buffer.
-- 
Andriy Gapon


More information about the zfs-devel mailing list