On 2019-05-26 11:49 a.m., Bart Van Assche wrote:
On 5/24/19 11:47 AM, Douglas Gilbert wrote:
This patchset restores some functionality that was in the v2 driver
version and has been broken by some external patch in the interim
period (20 years).
What is an "external patch"?
It's a patch made by someone with a first name like Christoph, Jens or
James (picking some names at random) that is applied whether or not
the maintainer of said driver approves ("ack's) or not.
IMO The maintainer should be able to restore features removed in this
fashion (I'm talking specifically about the neutering of
SG_FLAG_NO_DXFER) without review as it is in the documented interface.
Plus I know of at least one power user who was peeved to find out that
it was quietly broken.
Doug Gilbert
P.S the use case is mirrored disks: reading one disk for the actual data,
and the other disk to check if there is an IO error. So the second disk
doesn't need its data transferred to the user space (thus saving time).
