a stateid race and a cleanup
Subsystems (2)
v2 cover a stateid race and a cleanup
Cover letter this time explaining the v2: Anna helped me find that the first version's stable fix was wrong, but was fixed by the refactor patch that followed.
After putting in the logic to fix the stable version, it was messy enough that it made more sense to squash the two patches together. So, this time the first patch does rewrite nfs_need_update_open_stateid a bit more in order to handle both cases:
- where two OPENs race to NFS_OPEN_STATE and the second wins
- where an OPEN and CLOSE+1 race to update nfs4_state and CLOSE+1 wins
The end result is that these two patches are code-equivalent to the first three. (It is still getting one final run through my testing, but I haven't delayed posting for that).
Benjamin Coddington (2):
NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
NFSv4: cleanup unused zero_stateid copy
fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- fs/nfs/nfs4state.c | 8 ++------ 2 files changed, 19 insertions(+), 16 deletions(-)
Signature · 1 line
2.20.1
PATCH v2 1/2 NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Since commit 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID in CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE races with the update of the nfs_state:
Process 1 Process 2 Server
========= ========= ========
OPEN file
OPEN file
Reply OPEN (1)
Reply OPEN (2)
Update state (1)
CLOSE file (1)
Reply OLD_STATEID (1)
CLOSE file (2)
Reply CLOSE (-1)
Update state (2)
wait for state change
OPEN file
wake
CLOSE file
OPEN file
wake
CLOSE file
...
...As long as the first process continues updating state, the second process will fail to exit the loop in nfs_set_open_stateid_locked(). This livelock has been observed in generic/168.
Fix this by detecting the case in nfs_need_update_open_stateid() and then exit the loop if:
- the state is NFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateid
--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)
fs/nfs/nfs4proc.c +17 -10
@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state)static bool nfs_need_update_open_stateid (struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state->open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence number */ +if (state_matches_other &&nfs4_stateid_is_newer (stateid, &state->open_stateid)) { +nfs_state_log_out_of_order_open_stateid (state, stateid); +return true; + } + }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {nfs_state_log_update_open_stateid (state); -else +return true; + } +if (!state_matches_other && !seqid_one) {set_bit (NFS_STATE_CHANGE_WAIT , &state->flags); -return true; - } - -if (nfs4_stateid_is_newer (stateid, &state->open_stateid)) { -nfs_state_log_out_of_order_open_stateid (state, stateid); -return true; +return true; + } }return false; }
Signature · 1 line
2.20.1
Re: [PATCH 1/2 v2] NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Quoted 70 lines · Benjamin Coddington
Since commit
0e0cb35b417f("NFSv4: HandleNFS4ERR_OLD_STATEIDin CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE races
with the update of thenfs_state:Process 1 Process 2 Server ========= ========= ======== OPEN file OPEN file Reply OPEN (1) Reply OPEN (2) Update state (1) CLOSE file (1) Reply OLD_STATEID (1) CLOSE file (2) Reply CLOSE (-1) Update state (2) wait for state change OPEN file wake CLOSE file OPEN file wake CLOSE file ... ...As long as the first process continues updating state, the second process
will fail to exit the loop innfs_set_open_stateid_locked(). This livelock
has been observed in generic/168.Fix this by detecting the case in
nfs_need_update_open_stateid()and then exit the loop if:
- the state isNFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateidFixes 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID inCLOSE/OPEN_DOWNGRADE")
Cc [email protected] # v5.4+Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)fs/nfs/nfs4proc.c +5
-3@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) static bool nfs_need_update_open_stateid(struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state->open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence number */
fs/nfs/nfs4proc.c +1
-0+ if (state_matches_other &&
nfs4_stateid_is_newer(stateid, &state->open_stateid)) {fs/nfs/nfs4proc.c +1
-0+ nfs_state_log_out_of_order_open_stateid(state,stateid);
fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1602
Quoted 5 lines
fs/nfs/nfs4proc.c +5
-0+return true; + } + }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {
Why are we looking at the value of state_matches_other here? If the NFS_OPEN_STATE flags is not set, then the state->open_stateid is not initialised, so how does it make sense to look at a comparison with the incoming stateid?
fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1606
Quoted 5 lines
fs/nfs/nfs4proc.c +3
-1nfs_state_log_update_open_stateid (state); - else +return true; + } +if (!state_matches_other && !seqid_one) {
Ditto.
Quoted 13 lines
fs/nfs/nfs4proc.c +0
-5set_bit (NFS_STATE_CHANGE_WAIT , &state->flags); -return true; - } - -if (nfs4_stateid_is_newer (stateid, &state->open_stateid)) { - nfs_state_log_out_of_order_open_stateid(state,stateid);
fs/nfs/nfs4proc.c +2
-1-return true; +return true; + } }return false; }
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH 1/2 v2] NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Quoted 73 lines · Trond Myklebust
Quoted 72 lines · Benjamin Coddington
Since commit
0e0cb35b417f("NFSv4: HandleNFS4ERR_OLD_STATEIDin CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE races
with the update of thenfs_state:Process 1 Process 2 Server ========= ========= ======== OPEN file OPEN file Reply OPEN (1) Reply OPEN (2) Update state (1) CLOSE file (1) Reply OLD_STATEID (1) CLOSE file (2) Reply CLOSE (-1) Update state (2) wait for state change OPEN file wake CLOSE file OPEN file wake CLOSE file ... ...As long as the first process continues updating state, the second process
will fail to exit the loop innfs_set_open_stateid_locked(). This livelock
has been observed in generic/168.Fix this by detecting the case in
nfs_need_update_open_stateid()and then exit the loop if:
- the state isNFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateidFixes 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID inCLOSE/OPEN_DOWNGRADE")
Cc [email protected] # v5.4+Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)fs/nfs/nfs4proc.c +5
-3@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) static bool nfs_need_update_open_stateid(struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state->open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence number */
fs/nfs/nfs4proc.c +1
-0+ if (state_matches_other &&
nfs4_stateid_is_newer(stateid, &state->open_stateid)) {fs/nfs/nfs4proc.c +1
-0+ nfs_state_log_out_of_order_open_stateid(state,stateid);
fs/nfs/nfs4proc.c +2
-0+return true; + }
Quoted 8 lines
fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1602
Quoted 3 lines
fs/nfs/nfs4proc.c +3
-0+ }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {Why are we looking at the value of
state_matches_otherhere? If theNFS_OPEN_STATEflags is not set, then thestate->open_stateidis not initialised, so how does it make sense to look at a comparison with the incoming stateid?
You're right - it doesn't make sense. I failed to clean it up out of my decision matrix. I'll send a v3 after it gets tested overnight.
Thanks for the look, and if you have another moment - what do you think about not bumping the seqid in the CLOSE/OPEN_DOWNGRADE path on OLD_STATEID if the state's refcount > 1? This would optimize away a lot of recovery for races like this, but I wonder if it would be possible to end up with two OPEN_DOWNGRADEs dueling that would never recover.
If you don't have the moment, no problem. Thanks again for looking at my patch.
Ben
Re: [PATCH 1/2 v2] NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Quoted 95 lines · Benjamin Coddington
Quoted 81 lines · Trond Myklebust
Quoted 74 lines · Benjamin Coddington
Since commit
0e0cb35b417f("NFSv4: HandleNFS4ERR_OLD_STATEIDin CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE
races
with the update of thenfs_state:Process 1 Process 2 Server ========= ========= ======== OPEN file OPEN file Reply OPEN (1) Reply OPEN (2) Update state (1) CLOSE file (1) Reply OLD_STATEID (1) CLOSE file (2) Reply CLOSE (-1) Update state (2) wait for state change OPEN file wake CLOSE file OPEN file wake CLOSE file ... ...As long as the first process continues updating state, the second process
will fail to exit the loop innfs_set_open_stateid_locked(). This
livelock
has been observed in generic/168.Fix this by detecting the case in
nfs_need_update_open_stateid()and
then exit the loop if:
- the state isNFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateidFixes 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID inCLOSE/OPEN_DOWNGRADE")
Cc [email protected] # v5.4+Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)fs/nfs/nfs4proc.c +5
-3@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) static bool nfs_need_update_open_stateid(struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state->open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence number */
fs/nfs/nfs4proc.c +1
-0+ if (state_matches_other &&
nfs4_stateid_is_newer(stateid, &state->open_stateid)) {fs/nfs/nfs4proc.c +1
-0+ nfs_state_log_out_of_order_open_stateid(state,stateid);
fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1602
Quoted 5 lines
fs/nfs/nfs4proc.c +5
-0+return true; + } + }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {Quoted 7 lines
Why are we looking at the value of
state_matches_otherhere? If theNFS_OPEN_STATEflags is not set, then thestate->open_stateidis not
initialised, so how does it make sense to look at a comparison with
the
incoming stateid?You're right - it doesn't make sense. I failed to clean it up out of my
decision matrix. I'll send a v3 after it gets tested overnight.
Quoted 11 lines
Thanks for the look, and if you have another moment - what do you think
about not bumping the seqid in the CLOSE/OPEN_DOWNGRADE path onOLD_STATEID
if the state's refcount > 1? This would optimize away a lot of recovery
for
races like this, but I wonder if it would be possible to end up with twoOPEN_DOWNGRADEsdueling that would never recover.
It would lead to a return of the infinite loop in cases where the client misses a reply to an OPEN or CLOSE due to a soft timeout (which is an important use case for us).
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH 1/2 v2] NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Quoted 118 lines · Trond Myklebust
Quoted 104 lines · Benjamin Coddington
Quoted 88 lines · Trond Myklebust
Quoted 81 lines · Benjamin Coddington
Since commit
0e0cb35b417f("NFSv4: HandleNFS4ERR_OLD_STATEIDin
CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE
races
with the update of thenfs_state:Process 1 Process 2 Server ========= ========= ======== OPEN file OPEN file Reply OPEN (1) Reply OPEN (2) Update state (1) CLOSE file (1) Reply OLD_STATEID (1) CLOSE file (2) Reply CLOSE (-1) Update state (2) wait for state change OPEN file wake CLOSE file OPEN file wake CLOSE file ... ...As long as the first process continues updating state, the second
process
will fail to exit the loop innfs_set_open_stateid_locked(). This
livelock
has been observed in generic/168.Fix this by detecting the case in
nfs_need_update_open_stateid()
and
then exit the loop if:
- the state isNFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateidFixes 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID inCLOSE/OPEN_DOWNGRADE")
Cc [email protected] # v5.4+Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)fs/nfs/nfs4proc.c +5
-3@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) static bool nfs_need_update_open_stateid(struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state- >open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence
number */fs/nfs/nfs4proc.c +1
-0+ if (state_matches_other &&
nfs4_stateid_is_newer(stateid, &state->open_stateid)) {fs/nfs/nfs4proc.c +1
-0+ nfs_state_log_out_of_order_open_stateid(state,
stateid);fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1602
Quoted 5 lines
fs/nfs/nfs4proc.c +5
-0+return true; + } + }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {Quoted 8 lines
Why are we looking at the value of
state_matches_otherhere? If theNFS_OPEN_STATEflags is not set, then thestate->open_stateidis not
initialised, so how does it make sense to look at a comparison with
the
incoming stateid?You're right - it doesn't make sense. I failed to clean it up out of
my
decision matrix. I'll send a v3 after it gets tested overnight.Quoted 12 lines
Thanks for the look, and if you have another moment - what do you think
about not bumping the seqid in the CLOSE/OPEN_DOWNGRADE path onOLD_STATEID
if the state's refcount > 1? This would optimize away a lot of recovery
for
races like this, but I wonder if it would be possible to end up with
twoOPEN_DOWNGRADEsdueling that would never recover.
Quoted 5 lines
It would lead to a return of the infinite loop in cases where the client misses a reply to an OPEN or CLOSE due to a soft timeout (which
is an important use case for us).
In principle, RFC5661 says that the client is supposed to process all stateid operations in the order dictated by the seqid. The one problem with that mandate is that open-by-filename doesn't allow us to know whether or not a seqid bump is outstanding. This is why we have the 5 second timeout wait in nfs_set_open_stateid_locked().
We could do something similar to that for CLOSE by setting the seqid to 0, and then ensuring we process the CLOSE in the order the server ends up processing it. Unfortunately, that would require us to replay any OPEN calls that got shut down because we used seqid 0 (it would also not work for NFSv4.0)... So perhaps the best solution would be to instead allow CLOSE to wait for outstanding OPENs to complete, just like we do in nfs_set_open_stateid_locked()? Thoughts?
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH 1/2 v2] NFSv4: Fix a livelock when CLOSE pre-emptively bumps state sequence
Quoted 119 lines · Trond Myklebust
Quoted 118 lines · Trond Myklebust
Quoted 116 lines · Benjamin Coddington
Quoted 88 lines · Trond Myklebust
Quoted 81 lines · Benjamin Coddington
Since commit
0e0cb35b417f("NFSv4: HandleNFS4ERR_OLD_STATEIDin
CLOSE/OPEN_DOWNGRADE") the following livelock may occur if a CLOSE
races
with the update of thenfs_state:Process 1 Process 2 Server ========= ========= ======== OPEN file OPEN file Reply OPEN (1) Reply OPEN (2) Update state (1) CLOSE file (1) Reply OLD_STATEID (1) CLOSE file (2) Reply CLOSE (-1) Update state (2) wait for state change OPEN file wake CLOSE file OPEN file wake CLOSE file ... ...As long as the first process continues updating state, the second
process
will fail to exit the loop innfs_set_open_stateid_locked(). This
livelock
has been observed in generic/168.Fix this by detecting the case in
nfs_need_update_open_stateid()
and
then exit the loop if:
- the state isNFS_OPEN_STATE, and
- the stateid sequence is > 1, and
- the stateid doesn't match the current open stateidFixes 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID inCLOSE/OPEN_DOWNGRADE")
Cc [email protected] # v5.4+Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/nfs4proc.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-)fs/nfs/nfs4proc.c +5
-3@@ -1588,18 +1588,25 @@ static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) static bool nfs_need_update_open_stateid(struct nfs4_state *state,const nfs4_stateid *stateid) { -if (test_bit (NFS_OPEN_STATE , &state->flags) ==0 || - !nfs4_stateid_match_other (stateid, &state- > open_stateid)) { -if (stateid->seqid ==cpu_to_be32 (1 )) +bool state_matches_other =nfs4_stateid_match_other (stateid, &state->open_stateid); +bool seqid_one = stateid->seqid ==cpu_to_be32 (1 ); + +if (test_bit (NFS_OPEN_STATE , &state->flags)) { +/* The common case - we're updating to a new sequence
number */fs/nfs/nfs4proc.c +1
-0+ if (state_matches_other &&
nfs4_stateid_is_newer(stateid, &state->open_stateid)) {fs/nfs/nfs4proc.c +1
-0+ nfs_state_log_out_of_order_open_stateid(state,
stateid);fs/nfs/nfs4proc.c static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state) · line 1602
Quoted 5 lines
fs/nfs/nfs4proc.c +5
-0+return true; + } + }else { +/* This is the first OPEN */ +if (!state_matches_other && seqid_one) {Quoted 8 lines
Why are we looking at the value of
state_matches_otherhere? If theNFS_OPEN_STATEflags is not set, then thestate->open_stateidis not
initialised, so how does it make sense to look at a comparison with
the
incoming stateid?You're right - it doesn't make sense. I failed to clean it up out of
my
decision matrix. I'll send a v3 after it gets tested overnight.Thanks for the look, and if you have another moment - what do you think
about not bumping the seqid in the CLOSE/OPEN_DOWNGRADE path onOLD_STATEID
if the state's refcount > 1? This would optimize away a lot of recovery
for
races like this, but I wonder if it would be possible to end up with
twoOPEN_DOWNGRADEsdueling that would never recover.
Quoted 4 lines
Quoted 4 lines
It would lead to a return of the infinite loop in cases where the client misses a reply to an OPEN or CLOSE due to a soft timeout (which
is an important use case for us).
I hadn't thought about that, but doesn't the task return and decrement the refcount on nfs4_state?
Quoted 6 lines
In principle, RFC5661 says that the client is supposed to process all stateid operations in the order dictated by the seqid. The one problem with that mandate is that open-by-filename doesn't allow us to know whether or not a seqid bump is outstanding. This is why we have the 5 second timeout wait in
nfs_set_open_stateid_locked().
Quoted 7 lines
We could do something similar to that for CLOSE by setting the seqid to 0, and then ensuring we process the CLOSE in the order the server ends up processing it. Unfortunately, that would require us to replay any OPEN calls that got shut down because we used seqid 0 (it would also not work for NFSv4.0)... So perhaps the best solution would be to instead allow CLOSE to wait for outstanding OPENs to complete, just like we do in
nfs_set_open_stateid_locked()? Thoughts?
How can you know if there are outstanding OPENs? I thought that was the whole intention of the CLOSE -> OLD_STATEID retry bump -- you cannot know if the server was able to process an interrupted OPEN.
I had a dumb patch that just caused CLOSE to delay a bit before retrying if it received OLD_STATEID, but I figured that was vulnerable to another livelock where another process could just keep doing open(), close() as well.
Ben
PATCH v2 2/2 NFSv4: cleanup unused zero_stateid copy
In reply to Benjamin CoddingtonSince commit d9aba2b40de6 ("NFSv4: Don't use the zero stateid with layoutget") the zero stateid will never be used.
--- fs/nfs/nfs4state.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)
fs/nfs/nfs4state.c +2 -6
@@ -1018,18 +1018,14 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst,bool nfs4_copy_open_stateid (nfs4_stateid *dst,struct nfs4_state *state) {bool ret; -const nfs4_stateid *src;int seq;do { ret = false; - src = &zero_stateid; seq =read_seqbegin (&state->seqlock); -if (test_bit (NFS_OPEN_STATE , &state->flags)) { - src = &state->open_stateid; +if (test_bit (NFS_OPEN_STATE , &state->flags)) ret = true; - } -nfs4_stateid_copy (dst, src); +nfs4_stateid_copy (dst, &state->open_stateid); }while (read_seqretry (&state->seqlock, seq));return ret; }
Signature · 1 line
2.20.1