| Message ID | 1393959943-9919-4-git-send-email-trond.myklebust@primarydata.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, Mar 4, 2014 at 2:05 PM, Trond Myklebust <trond.myklebust@primarydata.com> wrote: > If the open stateid could not be recovered, or the file locks were lost, > then we should fail the truncate() operation altogether. > > Reported-by: Andy Adamson <andros@netapp.com> > Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com > Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> > --- > fs/nfs/nfs4proc.c | 9 ++++++--- > 1 file changed, 6 insertions(+), 3 deletions(-) > > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c > index 44e088dc357c..daf41182ecfb 100644 > --- a/fs/nfs/nfs4proc.c > +++ b/fs/nfs/nfs4proc.c > @@ -2398,13 +2398,16 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, > > if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) { > /* Use that stateid */ > - } else if (truncate && state != NULL && nfs4_valid_open_stateid(state)) { > + } else if (truncate && state != NULL) { > struct nfs_lockowner lockowner = { > .l_owner = current->files, > .l_pid = current->tgid, > }; > - nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, > - &lockowner); > + if (!nfs4_valid_open_stateid(state)) > + return -EBADF; > + if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, > + &lockowner) < 0) AFICS this means -EIO = lost lock. Why fail the setattr? the file handle is not bad. Why not send a ZERO_STATEID? At any rate, -EBADF looks wrong to me. -->Andy > + return -EBADF; > } else > nfs4_stateid_copy(&arg.stateid, &zero_stateid); > > -- > 1.8.5.3 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-nfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mar 5, 2014, at 6:40, Andy Adamson <androsadamson@gmail.com> wrote: > On Tue, Mar 4, 2014 at 2:05 PM, Trond Myklebust > <trond.myklebust@primarydata.com> wrote: >> If the open stateid could not be recovered, or the file locks were lost, >> then we should fail the truncate() operation altogether. >> >> Reported-by: Andy Adamson <andros@netapp.com> >> Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com >> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> >> --- >> fs/nfs/nfs4proc.c | 9 ++++++--- >> 1 file changed, 6 insertions(+), 3 deletions(-) >> >> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c >> index 44e088dc357c..daf41182ecfb 100644 >> --- a/fs/nfs/nfs4proc.c >> +++ b/fs/nfs/nfs4proc.c >> @@ -2398,13 +2398,16 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, >> >> if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) { >> /* Use that stateid */ >> - } else if (truncate && state != NULL && nfs4_valid_open_stateid(state)) { >> + } else if (truncate && state != NULL) { >> struct nfs_lockowner lockowner = { >> .l_owner = current->files, >> .l_pid = current->tgid, >> }; >> - nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, >> - &lockowner); >> + if (!nfs4_valid_open_stateid(state)) >> + return -EBADF; >> + if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, >> + &lockowner) < 0) > > AFICS this means -EIO = lost lock. Why fail the setattr? the file > handle is not bad. Why not send a ZERO_STATEID? At any rate, -EBADF > looks wrong to me. What should we use instead of EBADF? EIO doesn’t really help you figure out what is wrong or how to fix it.
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 44e088dc357c..daf41182ecfb 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2398,13 +2398,16 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) { /* Use that stateid */ - } else if (truncate && state != NULL && nfs4_valid_open_stateid(state)) { + } else if (truncate && state != NULL) { struct nfs_lockowner lockowner = { .l_owner = current->files, .l_pid = current->tgid, }; - nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, - &lockowner); + if (!nfs4_valid_open_stateid(state)) + return -EBADF; + if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, + &lockowner) < 0) + return -EBADF; } else nfs4_stateid_copy(&arg.stateid, &zero_stateid);
If the open stateid could not be recovered, or the file locks were lost, then we should fail the truncate() operation altogether. Reported-by: Andy Adamson <andros@netapp.com> Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> --- fs/nfs/nfs4proc.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)