On 11 Sep 2022, at 15:39, Al Viro wrote: >> >>> + for (int i = sd->len; i > 0; i -= PAGE_SIZE) >>> + svc_rqst_replace_page(rqstp, page++); >>> + if (rqstp->rq_res.page_len == 0) // first call >>> + rqstp->rq_res.page_base = offset % PAGE_SIZE; >>> rqstp->rq_res.page_len += sd->len; >>> return sd->len; >>> } >> >> I could take this through the nfsd for-rc tree, but that's based >> on 5.19-rc7 so it doesn't have f0f6b614f83d. I don't think will >> break functionality, but I'm wondering if it would be better for >> you to take this through your tree to improve bisect-ability. >> >> If you agree and Ben reports a Tested-by:, then here's my >> >> Acked-by: Chuck Lever <chuck.lever@xxxxxxxxxx> > > OK, I'll wait for Tested-by and send it to Linus. No new crashes or corruption after several passes through xfstests. I've also tested the RDMA transport on soft iwarp. Thanks for the fix here. Tested-by: Benjamin Coddington <bcodding@xxxxxxxxxx> Ben
