On Mon, 2019-08-26 at 19:12 -0400, Jan Stancek wrote: > ----- Original Message ----- > > On Mon, 2019-08-26 at 10:38 -0400, Jan Stancek wrote: > > > ----- Original Message ----- > > > > Hi Jan and Cyril, > > > > > > > > On Mon, 26 Aug 2019 at 16:35, Jan Stancek <jstancek@xxxxxxxxxx> > > > > wrote: > > > > > > > > > > ----- Original Message ----- > > > > > > Hi! > > > > > > > Do you see this LTP prot_hsymlinks failure on linux next > > > > > > > 20190823 on > > > > > > > x86_64 and i386 devices? > > > > > > > > > > > > > > test output log, > > > > > > > useradd: failure while writing changes to /etc/passwd > > > > > > > useradd: /home/hsym was created, but could not be removed > > > > > > > > > > > > This looks like an unrelated problem, failure to write to > > > > > > /etc/passwd > > > > > > probably means that filesystem is full or some problem > > > > > > happend > > > > > > and how > > > > > > is remounted RO. > > > > > > > > > > In Naresh' example, root is on NFS: > > > > > root=/dev/nfs rw > > > > > > > > > > nfsroot=10.66.16.123:/var/lib/lava/dispatcher/tmp/886412/extr > > > > > act- > > > > > nfsrootfs-tyuevoxm,tcp,hard,intr > > > > > > > > Right ! > > > > root is mounted on NFS. > > > > > > > > > 10.66.16.123:/var/lib/lava/dispatcher/tmp/886412/extract- > > > > > nfsrootfs-tyuevoxm > > > > > on / type nfs > > > > > (rw,relatime,vers=2,rsize=4096,wsize=4096,namlen=255,hard,nol > > > > > ock, > > > > > proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.66.16.123, > > > > > moun > > > > > tvers=1,mountproto=tcp,local_lock=all,addr=10.66.16.123) > > > > > devtmpfs on /dev type devtmpfs > > > > > (rw,relatime,size=3977640k,nr_inodes=994410,mode=755) > > > > > > > > > The only thing I can think of that might cause an EIO on NFSv2 > > would be > > this patch > > http://git.linux-nfs.org/?p=trondmy/linux-nfs.git;a=commitdiff;h=627d48e597ec5993c4abb3b81dc75e554a07c7c0 > > assuming that a bind-related error is leaking through. > > > > I'd suggest something like the following to fix it up: > > No change with that patch, > but following one fixes it for me: > > diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c > index 20b3717cd7ca..56cefa0ab804 100644 > --- a/fs/nfs/pagelist.c > +++ b/fs/nfs/pagelist.c > @@ -590,7 +590,7 @@ static void nfs_pgio_rpcsetup(struct > nfs_pgio_header *hdr, > } > > hdr->res.fattr = &hdr->fattr; > - hdr->res.count = 0; > + hdr->res.count = count; > hdr->res.eof = 0; > hdr->res.verf = &hdr->verf; > nfs_fattr_init(&hdr->fattr); > > which is functionally revert of "NFS: Fix initialisation of I/O > result struct in nfs_pgio_rpcsetup". > > This hunk caught my eye, could res.eof == 0 explain those I/O errors? Interesting hypothesis. It could if res.count ends up being 0. So does the following also fix the problem? 8<---------------------------------------- >From b5bc0812350e94f8c9331174d22f24692411aef9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx> Date: Mon, 26 Aug 2019 20:41:16 -0400 Subject: [PATCH] NFSv2: Fix eof handling If we received a reply from the server with a zero length read and no error, then that implies we are at eof. Signed-off-by: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx> --- fs/nfs/proc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index 5552fa8b6e12..5919878549d2 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -594,7 +594,8 @@ static int nfs_read_done(struct rpc_task *task, struct nfs_pgio_header *hdr) /* Emulate the eof flag, which isn't normally needed in NFSv2 * as it is guaranteed to always return the file attributes */ - if (hdr->args.offset + hdr->res.count >= hdr->res.fattr->size) + if (hdr->res.count == 0 && hdr->args.count > 0 || + hdr->args.offset + hdr->res.count >= hdr->res.fattr->size) hdr->res.eof = 1; } return 0; -- 2.21.0 -- Trond Myklebust Linux NFS client maintainer, Hammerspace trond.myklebust@xxxxxxxxxxxxxxx
