* [PATCH 1/1] pnfs_submit: Restore forgetful mode io draining @ 2010-10-07 9:35 Fred Isaman 2010-10-12 18:04 ` Benny Halevy 0 siblings, 1 reply; 3+ messages in thread From: Fred Isaman @ 2010-10-07 9:35 UTC (permalink / raw) To: linux-nfs Recent changes to the waitq code removed the wait for io to drain in the case of a "forgetful" return. Restore the previous behavior. Signed-off-by: Fred Isaman <iisaman@netapp.com> --- fs/nfs/nfs4proc.c | 6 +++++- fs/nfs/pnfs.c | 9 +++------ include/linux/nfs_xdr.h | 1 + 3 files changed, 9 insertions(+), 7 deletions(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index c854814..7727f592 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -5639,7 +5639,11 @@ nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata) rpc_sleep_on(&nfsi->lo_rpcwaitq, task, NULL); return; } - + if (lrp->stateid) { + /* Forget the layout, without sending the return */ + rpc_exit(task, 0); + return; + } if (nfs4_setup_sequence(server, NULL, &lrp->args.seq_args, &lrp->res.seq_res, 0, task)) return; diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index f573219..eb4dfdf 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -612,7 +612,7 @@ pnfs_return_layout_barrier(struct nfs_inode *nfsi, static int return_layout(struct inode *ino, struct pnfs_layout_range *range, enum pnfs_layoutreturn_type type, struct pnfs_layout_hdr *lo, - bool wait) + bool wait, const nfs4_stateid *stateid) { struct nfs4_layoutreturn *lrp; struct nfs_server *server = NFS_SERVER(ino); @@ -633,6 +633,7 @@ return_layout(struct inode *ino, struct pnfs_layout_range *range, lrp->args.return_type = type; lrp->args.range = *range; lrp->args.inode = ino; + lrp->stateid = stateid; status = nfs4_proc_layoutreturn(lrp, wait); out: @@ -688,11 +689,7 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, __func__, status); } } - - if (!stateid) - status = return_layout(ino, &arg, type, lo, wait); - else - pnfs_layoutreturn_release(lo, &arg); + status = return_layout(ino, &arg, type, lo, wait, stateid); } out: dprintk("<-- %s status: %d\n", __func__, status); diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index c72de21..80e6a36 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -277,6 +277,7 @@ struct nfs4_layoutreturn { struct nfs4_layoutreturn_args args; struct nfs4_layoutreturn_res res; struct rpc_cred *cred; + const nfs4_stateid *stateid; int rpc_status; }; -- 1.7.2.1 ^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH 1/1] pnfs_submit: Restore forgetful mode io draining 2010-10-07 9:35 [PATCH 1/1] pnfs_submit: Restore forgetful mode io draining Fred Isaman @ 2010-10-12 18:04 ` Benny Halevy 2010-10-12 18:48 ` Fred Isaman 0 siblings, 1 reply; 3+ messages in thread From: Benny Halevy @ 2010-10-12 18:04 UTC (permalink / raw) To: Fred Isaman; +Cc: linux-nfs On 2010-10-07 05:35, Fred Isaman wrote: > Recent changes to the waitq code removed the wait for io to drain > in the case of a "forgetful" return. Restore the previous behavior. > > Signed-off-by: Fred Isaman <iisaman@netapp.com> > --- > fs/nfs/nfs4proc.c | 6 +++++- > fs/nfs/pnfs.c | 9 +++------ > include/linux/nfs_xdr.h | 1 + > 3 files changed, 9 insertions(+), 7 deletions(-) > > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c > index c854814..7727f592 100644 > --- a/fs/nfs/nfs4proc.c > +++ b/fs/nfs/nfs4proc.c > @@ -5639,7 +5639,11 @@ nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata) > rpc_sleep_on(&nfsi->lo_rpcwaitq, task, NULL); > return; > } > - > + if (lrp->stateid) { > + /* Forget the layout, without sending the return */ > + rpc_exit(task, 0); > + return; > + } > if (nfs4_setup_sequence(server, NULL, &lrp->args.seq_args, > &lrp->res.seq_res, 0, task)) > return; > diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c > index f573219..eb4dfdf 100644 > --- a/fs/nfs/pnfs.c > +++ b/fs/nfs/pnfs.c > @@ -612,7 +612,7 @@ pnfs_return_layout_barrier(struct nfs_inode *nfsi, > static int > return_layout(struct inode *ino, struct pnfs_layout_range *range, > enum pnfs_layoutreturn_type type, struct pnfs_layout_hdr *lo, > - bool wait) > + bool wait, const nfs4_stateid *stateid) > { > struct nfs4_layoutreturn *lrp; > struct nfs_server *server = NFS_SERVER(ino); > @@ -633,6 +633,7 @@ return_layout(struct inode *ino, struct pnfs_layout_range *range, > lrp->args.return_type = type; > lrp->args.range = *range; > lrp->args.inode = ino; > + lrp->stateid = stateid; > > status = nfs4_proc_layoutreturn(lrp, wait); > out: > @@ -688,11 +689,7 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, > __func__, status); > } > } > - > - if (!stateid) > - status = return_layout(ino, &arg, type, lo, wait); > - else > - pnfs_layoutreturn_release(lo, &arg); > + status = return_layout(ino, &arg, type, lo, wait, stateid); > } > out: > dprintk("<-- %s status: %d\n", __func__, status); > diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h > index c72de21..80e6a36 100644 > --- a/include/linux/nfs_xdr.h > +++ b/include/linux/nfs_xdr.h > @@ -277,6 +277,7 @@ struct nfs4_layoutreturn { > struct nfs4_layoutreturn_args args; > struct nfs4_layoutreturn_res res; > struct rpc_cred *cred; > + const nfs4_stateid *stateid; > int rpc_status; > }; > So after reverting this patch in post-submit there's no real reason for passing the stateid to pnfs_return_layout, is it? I'm planning to do the following in the pnfs branch (post revert of your patch): git diff --stat -p -M fs/nfs/callback_proc.c | 5 ++--- fs/nfs/inode.c | 2 +- fs/nfs/nfs4state.c | 2 +- fs/nfs/pnfs.c | 11 +---------- fs/nfs/pnfs.h | 5 +---- 5 files changed, 6 insertions(+), 19 deletions(-) diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 6b560ce..93a9b2e 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -234,8 +234,7 @@ static int pnfs_recall_layout(void *data) if (pnfs_is_next_layout_stateid(NFS_I(inode)->layout, rl.cbl_stateid)) status = pnfs_return_layout(inode, &rl.cbl_seg, - &rl.cbl_stateid, RETURN_FILE, - false); + RETURN_FILE, false); else status = cpu_to_be32(NFS4ERR_DELAY); if (status) @@ -255,7 +254,7 @@ static int pnfs_recall_layout(void *data) /* IMPROVEME: This loop is inefficient, running in O(|s_inodes|^2) */ while ((ino = nfs_layoutrecall_find_inode(clp, &rl)) != NULL) { /* FIXME: need to check status on pnfs_return_layout */ - pnfs_return_layout(ino, &rl.cbl_seg, NULL, RETURN_FILE, false); + pnfs_return_layout(ino, &rl.cbl_seg, RETURN_FILE, false); iput(ino); } diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 437d9a6..cbd5b3f 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1418,7 +1418,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) */ void nfs4_evict_inode(struct inode *inode) { - pnfs_return_layout(inode, NULL, NULL, RETURN_FILE, true); + pnfs_return_layout(inode, NULL, RETURN_FILE, true); truncate_inode_pages(&inode->i_data, 0); end_writeback(inode); pnfs_destroy_layout(NFS_I(inode)); diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 3168d77..8d32f2f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -612,7 +612,7 @@ static void __nfs4_close(struct path *path, struct nfs4_state *state, .length = NFS4_MAX_UINT64, }; - pnfs_return_layout(state->inode, &range, NULL, + pnfs_return_layout(state->inode, &range, RETURN_FILE, wait); } diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index f573219..74d8532 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -642,7 +642,6 @@ out: int _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, - const nfs4_stateid *stateid, /* optional */ enum pnfs_layoutreturn_type type, bool wait) { @@ -675,11 +674,6 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, spin_unlock(&ino->i_lock); if (layoutcommit_needed(nfsi)) { - if (stateid && !wait) { /* callback */ - dprintk("%s: layoutcommit pending\n", __func__); - status = -EAGAIN; - goto out_put; - } status = pnfs_layoutcommit_inode(ino, wait); if (status) { /* Return layout even if layoutcommit fails */ @@ -689,10 +683,7 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, } } - if (!stateid) - status = return_layout(ino, &arg, type, lo, wait); - else - pnfs_layoutreturn_release(lo, &arg); + status = return_layout(ino, &arg, type, lo, wait); } out: dprintk("<-- %s status: %d\n", __func__, status); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 235709e..7ca22a0 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -184,7 +184,6 @@ pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, enum pnfs_iomode access_type); bool pnfs_return_layout_barrier(struct nfs_inode *, struct pnfs_layout_range *); int _pnfs_return_layout(struct inode *, struct pnfs_layout_range *, - const nfs4_stateid *stateid, /* optional */ enum pnfs_layoutreturn_type, bool wait); void set_pnfs_layoutdriver(struct nfs_server *, u32 id); void unset_pnfs_layoutdriver(struct nfs_server *); @@ -252,7 +251,6 @@ pnfs_layout_roc_iomode(struct nfs_inode *nfsi) static inline int pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, - const nfs4_stateid *stateid, /* optional */ enum pnfs_layoutreturn_type type, bool wait) { @@ -261,7 +259,7 @@ static inline int pnfs_return_layout(struct inode *ino, if (pnfs_enabled_sb(nfss) && (type != RETURN_FILE || has_layout(nfsi))) - return _pnfs_return_layout(ino, range, stateid, type, wait); + return _pnfs_return_layout(ino, range, type, wait); return 0; } @@ -350,7 +348,6 @@ pnfs_layout_roc_iomode(struct nfs_inode *nfsi) static inline int pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, - const nfs4_stateid *stateid, /* optional */ enum pnfs_layoutreturn_type type, bool wait) { ^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH 1/1] pnfs_submit: Restore forgetful mode io draining 2010-10-12 18:04 ` Benny Halevy @ 2010-10-12 18:48 ` Fred Isaman 0 siblings, 0 replies; 3+ messages in thread From: Fred Isaman @ 2010-10-12 18:48 UTC (permalink / raw) To: Benny Halevy; +Cc: linux-nfs On Tue, Oct 12, 2010 at 2:04 PM, Benny Halevy <bhalevy@panasas.com> wrote: > On 2010-10-07 05:35, Fred Isaman wrote: >> Recent changes to the waitq code removed the wait for io to drain >> in the case of a "forgetful" return. Restore the previous behavior. >> >> Signed-off-by: Fred Isaman <iisaman@netapp.com> >> --- >> fs/nfs/nfs4proc.c | 6 +++++- >> fs/nfs/pnfs.c | 9 +++------ >> include/linux/nfs_xdr.h | 1 + >> 3 files changed, 9 insertions(+), 7 deletions(-) >> >> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c >> index c854814..7727f592 100644 >> --- a/fs/nfs/nfs4proc.c >> +++ b/fs/nfs/nfs4proc.c >> @@ -5639,7 +5639,11 @@ nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata) >> rpc_sleep_on(&nfsi->lo_rpcwaitq, task, NULL); >> return; >> } >> - >> + if (lrp->stateid) { >> + /* Forget the layout, without sending the return */ >> + rpc_exit(task, 0); >> + return; >> + } >> if (nfs4_setup_sequence(server, NULL, &lrp->args.seq_args, >> &lrp->res.seq_res, 0, task)) >> return; >> diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c >> index f573219..eb4dfdf 100644 >> --- a/fs/nfs/pnfs.c >> +++ b/fs/nfs/pnfs.c >> @@ -612,7 +612,7 @@ pnfs_return_layout_barrier(struct nfs_inode *nfsi, >> static int >> return_layout(struct inode *ino, struct pnfs_layout_range *range, >> enum pnfs_layoutreturn_type type, struct pnfs_layout_hdr *lo, >> - bool wait) >> + bool wait, const nfs4_stateid *stateid) >> { >> struct nfs4_layoutreturn *lrp; >> struct nfs_server *server = NFS_SERVER(ino); >> @@ -633,6 +633,7 @@ return_layout(struct inode *ino, struct pnfs_layout_range *range, >> lrp->args.return_type = type; >> lrp->args.range = *range; >> lrp->args.inode = ino; >> + lrp->stateid = stateid; >> >> status = nfs4_proc_layoutreturn(lrp, wait); >> out: >> @@ -688,11 +689,7 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, >> __func__, status); >> } >> } >> - >> - if (!stateid) >> - status = return_layout(ino, &arg, type, lo, wait); >> - else >> - pnfs_layoutreturn_release(lo, &arg); >> + status = return_layout(ino, &arg, type, lo, wait, stateid); >> } >> out: >> dprintk("<-- %s status: %d\n", __func__, status); >> diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h >> index c72de21..80e6a36 100644 >> --- a/include/linux/nfs_xdr.h >> +++ b/include/linux/nfs_xdr.h >> @@ -277,6 +277,7 @@ struct nfs4_layoutreturn { >> struct nfs4_layoutreturn_args args; >> struct nfs4_layoutreturn_res res; >> struct rpc_cred *cred; >> + const nfs4_stateid *stateid; >> int rpc_status; >> }; >> > > So after reverting this patch in post-submit there's no real reason > for passing the stateid to pnfs_return_layout, is it? > I'm planning to do the following in the pnfs branch (post revert of your patch): It was used (merely as a boolean flag) to note that the return should be "forgotten". My understanding of your suggested patch is that the client would now always call LAYOUTRETURN, but without adding any of the race detection code we intentionally avoided writing because we were using the forgetful model. Fred > > git diff --stat -p -M > fs/nfs/callback_proc.c | 5 ++--- > fs/nfs/inode.c | 2 +- > fs/nfs/nfs4state.c | 2 +- > fs/nfs/pnfs.c | 11 +---------- > fs/nfs/pnfs.h | 5 +---- > 5 files changed, 6 insertions(+), 19 deletions(-) > > diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c > index 6b560ce..93a9b2e 100644 > --- a/fs/nfs/callback_proc.c > +++ b/fs/nfs/callback_proc.c > @@ -234,8 +234,7 @@ static int pnfs_recall_layout(void *data) > if (pnfs_is_next_layout_stateid(NFS_I(inode)->layout, > rl.cbl_stateid)) > status = pnfs_return_layout(inode, &rl.cbl_seg, > - &rl.cbl_stateid, RETURN_FILE, > - false); > + RETURN_FILE, false); > else > status = cpu_to_be32(NFS4ERR_DELAY); > if (status) > @@ -255,7 +254,7 @@ static int pnfs_recall_layout(void *data) > /* IMPROVEME: This loop is inefficient, running in O(|s_inodes|^2) */ > while ((ino = nfs_layoutrecall_find_inode(clp, &rl)) != NULL) { > /* FIXME: need to check status on pnfs_return_layout */ > - pnfs_return_layout(ino, &rl.cbl_seg, NULL, RETURN_FILE, false); > + pnfs_return_layout(ino, &rl.cbl_seg, RETURN_FILE, false); > iput(ino); > } > > diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c > index 437d9a6..cbd5b3f 100644 > --- a/fs/nfs/inode.c > +++ b/fs/nfs/inode.c > @@ -1418,7 +1418,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) > */ > void nfs4_evict_inode(struct inode *inode) > { > - pnfs_return_layout(inode, NULL, NULL, RETURN_FILE, true); > + pnfs_return_layout(inode, NULL, RETURN_FILE, true); > truncate_inode_pages(&inode->i_data, 0); > end_writeback(inode); > pnfs_destroy_layout(NFS_I(inode)); > diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c > index 3168d77..8d32f2f 100644 > --- a/fs/nfs/nfs4state.c > +++ b/fs/nfs/nfs4state.c > @@ -612,7 +612,7 @@ static void __nfs4_close(struct path *path, struct nfs4_state *state, > .length = NFS4_MAX_UINT64, > }; > > - pnfs_return_layout(state->inode, &range, NULL, > + pnfs_return_layout(state->inode, &range, > RETURN_FILE, wait); > } > > diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c > index f573219..74d8532 100644 > --- a/fs/nfs/pnfs.c > +++ b/fs/nfs/pnfs.c > @@ -642,7 +642,6 @@ out: > > int > _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, > - const nfs4_stateid *stateid, /* optional */ > enum pnfs_layoutreturn_type type, > bool wait) > { > @@ -675,11 +674,6 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, > spin_unlock(&ino->i_lock); > > if (layoutcommit_needed(nfsi)) { > - if (stateid && !wait) { /* callback */ > - dprintk("%s: layoutcommit pending\n", __func__); > - status = -EAGAIN; > - goto out_put; > - } > status = pnfs_layoutcommit_inode(ino, wait); > if (status) { > /* Return layout even if layoutcommit fails */ > @@ -689,10 +683,7 @@ _pnfs_return_layout(struct inode *ino, struct pnfs_layout_range *range, > } > } > > - if (!stateid) > - status = return_layout(ino, &arg, type, lo, wait); > - else > - pnfs_layoutreturn_release(lo, &arg); > + status = return_layout(ino, &arg, type, lo, wait); > } > out: > dprintk("<-- %s status: %d\n", __func__, status); > diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h > index 235709e..7ca22a0 100644 > --- a/fs/nfs/pnfs.h > +++ b/fs/nfs/pnfs.h > @@ -184,7 +184,6 @@ pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, > enum pnfs_iomode access_type); > bool pnfs_return_layout_barrier(struct nfs_inode *, struct pnfs_layout_range *); > int _pnfs_return_layout(struct inode *, struct pnfs_layout_range *, > - const nfs4_stateid *stateid, /* optional */ > enum pnfs_layoutreturn_type, bool wait); > void set_pnfs_layoutdriver(struct nfs_server *, u32 id); > void unset_pnfs_layoutdriver(struct nfs_server *); > @@ -252,7 +251,6 @@ pnfs_layout_roc_iomode(struct nfs_inode *nfsi) > > static inline int pnfs_return_layout(struct inode *ino, > struct pnfs_layout_range *range, > - const nfs4_stateid *stateid, /* optional */ > enum pnfs_layoutreturn_type type, > bool wait) > { > @@ -261,7 +259,7 @@ static inline int pnfs_return_layout(struct inode *ino, > > if (pnfs_enabled_sb(nfss) && > (type != RETURN_FILE || has_layout(nfsi))) > - return _pnfs_return_layout(ino, range, stateid, type, wait); > + return _pnfs_return_layout(ino, range, type, wait); > > return 0; > } > @@ -350,7 +348,6 @@ pnfs_layout_roc_iomode(struct nfs_inode *nfsi) > > static inline int pnfs_return_layout(struct inode *ino, > struct pnfs_layout_range *range, > - const nfs4_stateid *stateid, /* optional */ > enum pnfs_layoutreturn_type type, > bool wait) > { > -- > 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 > ^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2010-10-12 18:48 UTC | newest] Thread overview: 3+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2010-10-07 9:35 [PATCH 1/1] pnfs_submit: Restore forgetful mode io draining Fred Isaman 2010-10-12 18:04 ` Benny Halevy 2010-10-12 18:48 ` Fred Isaman
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.