ipc,sem: fix locking in semctl_main
PATCH 1/1 ipc,sem: fix locking in semctl_main
Andrew, here is the patch I promised :)
Everywhere else in the code, we check sma->sem_perm.deleted under the semaphore array lock, so we should do teh same here. The easy fix is to simply not drop and re-take the lock, but keep it locked.
This patch can be folded into ipcsem-fine-grained-locking-for-semtimedop.patch
--- ipc/sem.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)
ipc/sem.c +2 -3
@@ -1243,10 +1243,9 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, err = -EIDRM ; goto out_free; } -sem_unlock (sma,-1 ); - } + }else +sem_lock (sma,NULL ,-1 ); -sem_lock (sma,NULL ,-1 );for (i =0 ; i < sma->sem_nsems; i++) sem_io[i] = sma->sem_base[i].semval;sem_unlock (sma,-1 );
Quoted 7 lines · Rik van Riel
Andrew, here is the patch I promised :)
Everywhere else in the code, we check
sma->sem_perm.deletedunder the semaphore array lock, so we should do teh same here. The easy fix is to simply not drop and re-take the lock, but keep it locked.
CCing Linus, Sedat and Emmanuel.
This looks very much like a patch I asked Sedat to test during the weekend, which according to him didn't solve his issue. The patch is correct in any case.
Quoted 3 lines
This patch can be folded into ipcsem-fine-grained-locking-for-semtimedop.patch
Signed-off-by Rik van Riel <[email protected]>
Quoted 21 lines
--- ipc/sem.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)ipc/sem.c +2
-3@@ -1243,10 +1243,9 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, err = -EIDRM ; goto out_free; } -sem_unlock (sma,-1 ); - } + }else +sem_lock (sma,NULL ,-1 ); -sem_lock (sma,NULL ,-1 );for (i =0 ; i < sma->sem_nsems; i++) sem_io[i] = sma->sem_base[i].semval;sem_unlock (sma,-1 );
Quoted 9 lines · Davidlohr Bueso
Quoted 7 lines · Rik van Riel
Andrew, here is the patch I promised :)
Everywhere else in the code, we check
sma->sem_perm.deletedunder the semaphore array lock, so we should do teh same here. The easy fix is to simply not drop and re-take the lock, but keep it locked.
Quoted 6 lines
CCing Linus, Sedat and Emmanuel.
This looks very much like a patch I asked Sedat to test during the weekend, which according to him didn't solve his issue. The patch is correct in any case.
Your patch looked a bit different (checked for ipc-lock already taken, both patches - original and refreshed againt -next attached).
- Sedat -
Quoted 23 lines
Quoted 3 lines
This patch can be folded into ipcsem-fine-grained-locking-for-semtimedop.patch
Signed-off-by Rik van Riel <[email protected]>Acked-by Davidlohr Bueso <[email protected]>Quoted 17 lines
--- ipc/sem.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)ipc/sem.c +2
-2@@ -1243,10 +1243,9 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, err = -EIDRM ; goto out_free; } -sem_unlock (sma,-1 ); - } + }else +sem_lock (sma,NULL ,-1 );
Quoted 6 lines
ipc/sem.c static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, · line 1251
Quoted 4 lines
ipc/sem.c +0
-1-sem_lock (sma,NULL ,-1 );for (i =0 ; i < sma->sem_nsems; i++) sem_io[i] = sma->sem_base[i].semval;sem_unlock (sma,-1 );
From: Davidlohr Bueso <[email protected]>
Subject: ipc: sem: Check if the ipc lock has been already taken
Original patch from [1].
[1] http://marc.info/?t=136631457900005&r=1&w=2
ipc/sem.c +3 -2
@@ -1243,10 +1243,11 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, err = -EIDRM ; goto out_free; } -sem_unlock (sma,-1 ); } -sem_lock (sma,NULL ,-1 ); +/* has the ipc lock already been taken? */ +if (nsems <=SEMMSL_FAST ) +sem_lock (sma,NULL ,-1 );for (i =0 ; i < sma->sem_nsems; i++) sem_io[i] = sma->sem_base[i].semval;sem_unlock (sma,-1 );
From: Davidlohr Bueso <[email protected]>
Subject: ipc, sem: Check if the ipc lock has been already taken
Original patch from [1].
[ dileks: v2: Adapted to fit next-20130419 plus 3 ipc-sem-next patches from [2] ]
[1] http://marc.info/?t=136631457900005&r=1&w=2
[2] http://marc.info/?l=linux-next&m=136640233601913&w=2
ipc/sem.c +3 -2
@@ -1259,10 +1259,11 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum, err = -EIDRM ; goto out_free; } -sem_unlock (sma,-1 ); } -sem_lock (sma,NULL ,-1 ); +/* has the ipc lock already been taken? */ +if (nsems <=SEMMSL_FAST ) +sem_lock (sma,NULL ,-1 );for (i =0 ; i < sma->sem_nsems; i++) sem_io[i] = sma->sem_base[i].semval;sem_unlock (sma,-1 );