On Mon, Oct 25, 2021 at 09:12:14AM -0700, Bart Van Assche wrote:
> On 10/21/21 11:20 PM, Christoph Hellwig wrote:
> > The HPB support added this merge window is fundanetally flawed as it
> > uses blk_insert_cloned_request to insert a cloned request onto the same
> > queue as the one that the original request came from, leading to all
> > kinds of issues in blk-mq accounting (in addition to this API being
> > a special case for dm-mpath that should not see other users).
>
> I do not agree with removing the UFS HPB code from the upstream kernel at
> this time.
>
> One of the HPB users promised to look into removing the
> blk_insert_cloned_request() call from the UFS HPB code. Shouldn't that
> person be given the chance to come up with a patch before removal of the UFS
> HPB code is considered?
>
> Additionally, removing the UFS HPB code from the upstream kernel won't
> affect UFS users much. As far as I know all HPB users use Android. UFS HPB
> is supported by Android 12 and will also be supported by Android 13. Whether
> or not this patch is goes upstream won't affect the Android kernel. I am not
> aware of any plans to make any changes in the Android kernel UFS HPB code if
> this patch would be integrated in the upstream kernel.
Under this line of reasoning, why would upstream take the code at all?
{sigh}
Is there a link to where the HPB developer said they would look into
this? Perhaps until that happens this should be marked as BROKEN?
thanks,
greg k-h
