[Cluster-devel] [PATCH AUTOSEL 5.6 20/41] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 19ebc6cd0f2b..d0eceaff3cea 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -645,9 +645,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.6 14/41] gfs2: move privileged user check to gfs2_quota_lock_check
From: Bob Peterson [ Upstream commit 4ed0c30811cb4d30ef89850b787a53a84d5d2bcb ] Before this patch, function gfs2_quota_lock checked if it was called from a privileged user, and if so, it bypassed the quota check: superuser can operate outside the quotas. That's the wrong place for the check because the lock/unlock functions are separate from the lock_check function, and you can do lock and unlock without actually checking the quotas. This patch moves the check to gfs2_quota_lock_check. Signed-off-by: Bob Peterson Signed-off-by: Andreas Gruenbacher Signed-off-by: Sasha Levin --- fs/gfs2/quota.c | 3 +-- fs/gfs2/quota.h | 3 ++- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index e9f93045eb01..832d44782f74 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1040,8 +1040,7 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) u32 x; int error = 0; - if (capable(CAP_SYS_RESOURCE) || - sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; error = gfs2_quota_hold(ip, uid, gid); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 765627d9a91e..fe68a91dc16f 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -44,7 +44,8 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, int ret; ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ - if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + if (capable(CAP_SYS_RESOURCE) || + sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; ret = gfs2_quota_lock(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE); if (ret) -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.6 16/41] gfs2: Grab glock reference sooner in gfs2_add_revoke
From: Andreas Gruenbacher [ Upstream commit f4e2f5e1a527ce58fc9f85145b03704779a3123e ] This patch rearranges gfs2_add_revoke so that the extra glock reference is added earlier on in the function to avoid races in which the glock is freed before the new reference is taken. Signed-off-by: Andreas Gruenbacher Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/log.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index 60d911e293e6..2674feda1d7a 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -603,13 +603,13 @@ void gfs2_add_revoke(struct gfs2_sbd *sdp, struct gfs2_bufdata *bd) struct buffer_head *bh = bd->bd_bh; struct gfs2_glock *gl = bd->bd_gl; + sdp->sd_log_num_revoke++; + if (atomic_inc_return(&gl->gl_revokes) == 1) + gfs2_glock_hold(gl); bh->b_private = NULL; bd->bd_blkno = bh->b_blocknr; gfs2_remove_from_ail(bd); /* drops ref on bh */ bd->bd_bh = NULL; - sdp->sd_log_num_revoke++; - if (atomic_inc_return(&gl->gl_revokes) == 1) - gfs2_glock_hold(gl); set_bit(GLF_LFLUSH, &gl->gl_flags); list_add(&bd->bd_list, &sdp->sd_log_revokes); } -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.6 15/41] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 832d44782f74..a80b638b582e 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1113,7 +1113,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_qadata->qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1150,7 +1150,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.4 16/32] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 21820a5b388f..0290a22ebccf 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -639,9 +639,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.4 11/32] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index cbee745169b8..91ca4920bf03 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1113,7 +1113,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_qadata->qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1150,7 +1150,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.4 10/32] gfs2: move privileged user check to gfs2_quota_lock_check
From: Bob Peterson [ Upstream commit 4ed0c30811cb4d30ef89850b787a53a84d5d2bcb ] Before this patch, function gfs2_quota_lock checked if it was called from a privileged user, and if so, it bypassed the quota check: superuser can operate outside the quotas. That's the wrong place for the check because the lock/unlock functions are separate from the lock_check function, and you can do lock and unlock without actually checking the quotas. This patch moves the check to gfs2_quota_lock_check. Signed-off-by: Bob Peterson Signed-off-by: Andreas Gruenbacher Signed-off-by: Sasha Levin --- fs/gfs2/quota.c | 3 +-- fs/gfs2/quota.h | 3 ++- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 7c016a082aa6..cbee745169b8 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1040,8 +1040,7 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) u32 x; int error = 0; - if (capable(CAP_SYS_RESOURCE) || - sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; error = gfs2_quota_hold(ip, uid, gid); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 765627d9a91e..fe68a91dc16f 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -44,7 +44,8 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, int ret; ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ - if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + if (capable(CAP_SYS_RESOURCE) || + sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; ret = gfs2_quota_lock(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE); if (ret) -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 5.4 12/32] gfs2: Grab glock reference sooner in gfs2_add_revoke
From: Andreas Gruenbacher [ Upstream commit f4e2f5e1a527ce58fc9f85145b03704779a3123e ] This patch rearranges gfs2_add_revoke so that the extra glock reference is added earlier on in the function to avoid races in which the glock is freed before the new reference is taken. Signed-off-by: Andreas Gruenbacher Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/log.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index 47bc27d4169e..110e5c4db819 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -598,13 +598,13 @@ void gfs2_add_revoke(struct gfs2_sbd *sdp, struct gfs2_bufdata *bd) struct buffer_head *bh = bd->bd_bh; struct gfs2_glock *gl = bd->bd_gl; + sdp->sd_log_num_revoke++; + if (atomic_inc_return(&gl->gl_revokes) == 1) + gfs2_glock_hold(gl); bh->b_private = NULL; bd->bd_blkno = bh->b_blocknr; gfs2_remove_from_ail(bd); /* drops ref on bh */ bd->bd_bh = NULL; - sdp->sd_log_num_revoke++; - if (atomic_inc_return(&gl->gl_revokes) == 1) - gfs2_glock_hold(gl); set_bit(GLF_LFLUSH, &gl->gl_flags); list_add(&bd->bd_list, &sdp->sd_log_revokes); } -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.14 08/13] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 1e2ff4b32c79..aea1ed0aebd0 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -636,9 +636,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.4 2/5] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index f80ffccb0316..1eb737c466dd 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -541,9 +541,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.19 10/19] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index dd0f9bc13164..ce47c8233612 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1116,7 +1116,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_qadata->qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1153,7 +1153,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.4 1/5] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 3a31226531ea..4af00ed4960a 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1080,7 +1080,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1117,7 +1117,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.9 5/8] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index adc1a97cfe96..efd44d5645d8 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -548,9 +548,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.19 11/19] Revert "gfs2: Don't demote a glock until its revokes are written"
From: Bob Peterson [ Upstream commit b14c94908b1b884276a6608dea3d0b1b510338b7 ] This reverts commit df5db5f9ee112e76b5202fbc331f990a0fc316d6. This patch fixes a regression: patch df5db5f9ee112 allowed function run_queue() to bypass its call to do_xmote() if revokes were queued for the glock. That's wrong because its call to do_xmote() is what is responsible for calling the go_sync() glops functions to sync both the ail list and any revokes queued for it. By bypassing the call, gfs2 could get into a stand-off where the glock could not be demoted until its revokes are written back, but the revokes would not be written back because do_xmote() was never called. It "sort of" works, however, because there are other mechanisms like the log flush daemon (logd) that can sync the ail items and revokes, if it deems it necessary. The problem is: without file system pressure, it might never deem it necessary. Signed-off-by: Bob Peterson Signed-off-by: Sasha Levin --- fs/gfs2/glock.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index f8a5eef3d014..ccdd8c821abd 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -636,9 +636,6 @@ __acquires(&gl->gl_lockref.lock) goto out_unlock; if (nonblock) goto out_sched; - smp_mb(); - if (atomic_read(&gl->gl_revokes) != 0) - goto out_sched; set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags); GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE); gl->gl_target = gl->gl_demote_state; -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.9 4/8] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index fb9b1d702351..fb2e0ad945bf 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1112,7 +1112,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_qadata->qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1149,7 +1149,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.14 06/13] gfs2: move privileged user check to gfs2_quota_lock_check
From: Bob Peterson [ Upstream commit 4ed0c30811cb4d30ef89850b787a53a84d5d2bcb ] Before this patch, function gfs2_quota_lock checked if it was called from a privileged user, and if so, it bypassed the quota check: superuser can operate outside the quotas. That's the wrong place for the check because the lock/unlock functions are separate from the lock_check function, and you can do lock and unlock without actually checking the quotas. This patch moves the check to gfs2_quota_lock_check. Signed-off-by: Bob Peterson Signed-off-by: Andreas Gruenbacher Signed-off-by: Sasha Levin --- fs/gfs2/quota.c | 3 +-- fs/gfs2/quota.h | 3 ++- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index e700fb162664..a833e2e07167 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1039,8 +1039,7 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) u32 x; int error = 0; - if (capable(CAP_SYS_RESOURCE) || - sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; error = gfs2_quota_hold(ip, uid, gid); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 836f29480be6..e3a6e2404d11 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -47,7 +47,8 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, int ret; ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ - if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + if (capable(CAP_SYS_RESOURCE) || + sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; ret = gfs2_quota_lock(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE); if (ret) -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.9 3/8] gfs2: move privileged user check to gfs2_quota_lock_check
From: Bob Peterson [ Upstream commit 4ed0c30811cb4d30ef89850b787a53a84d5d2bcb ] Before this patch, function gfs2_quota_lock checked if it was called from a privileged user, and if so, it bypassed the quota check: superuser can operate outside the quotas. That's the wrong place for the check because the lock/unlock functions are separate from the lock_check function, and you can do lock and unlock without actually checking the quotas. This patch moves the check to gfs2_quota_lock_check. Signed-off-by: Bob Peterson Signed-off-by: Andreas Gruenbacher Signed-off-by: Sasha Levin --- fs/gfs2/quota.c | 3 +-- fs/gfs2/quota.h | 3 ++- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index c2ca9566b764..fb9b1d702351 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1039,8 +1039,7 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) u32 x; int error = 0; - if (capable(CAP_SYS_RESOURCE) || - sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; error = gfs2_quota_hold(ip, uid, gid); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 836f29480be6..e3a6e2404d11 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -47,7 +47,8 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, int ret; ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ - if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + if (capable(CAP_SYS_RESOURCE) || + sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; ret = gfs2_quota_lock(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE); if (ret) -- 2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.14 07/13] gfs2: don't call quota_unhold if quotas are not locked
From: Bob Peterson
[ Upstream commit c9cb9e381985bbbe8acd2695bbe6bd24bf06b81c ]
Before this patch, function gfs2_quota_unlock checked if quotas are
turned off, and if so, it branched to label out, which called
gfs2_quota_unhold. With the new system of gfs2_qa_get and put, we
no longer want to call gfs2_quota_unhold or we won't balance our
gets and puts.
Signed-off-by: Bob Peterson
Signed-off-by: Andreas Gruenbacher
Signed-off-by: Sasha Levin
---
fs/gfs2/quota.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index a833e2e07167..d34a668a432f 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -1112,7 +1112,7 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
int found;
if (!test_and_clear_bit(GIF_QD_LOCKED, &ip->i_flags))
- goto out;
+ return;
for (x = 0; x < ip->i_qadata->qa_qd_num; x++) {
struct gfs2_quota_data *qd;
@@ -1149,7 +1149,6 @@ void gfs2_quota_unlock(struct gfs2_inode *ip)
qd_unlock(qda[x]);
}
-out:
gfs2_quota_unhold(ip);
}
--
2.25.1
[Cluster-devel] [PATCH AUTOSEL 4.19 09/19] gfs2: move privileged user check to gfs2_quota_lock_check
From: Bob Peterson [ Upstream commit 4ed0c30811cb4d30ef89850b787a53a84d5d2bcb ] Before this patch, function gfs2_quota_lock checked if it was called from a privileged user, and if so, it bypassed the quota check: superuser can operate outside the quotas. That's the wrong place for the check because the lock/unlock functions are separate from the lock_check function, and you can do lock and unlock without actually checking the quotas. This patch moves the check to gfs2_quota_lock_check. Signed-off-by: Bob Peterson Signed-off-by: Andreas Gruenbacher Signed-off-by: Sasha Levin --- fs/gfs2/quota.c | 3 +-- fs/gfs2/quota.h | 3 ++- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 0efae7a0ee80..dd0f9bc13164 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -1043,8 +1043,7 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid) u32 x; int error = 0; - if (capable(CAP_SYS_RESOURCE) || - sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) return 0; error = gfs2_quota_hold(ip, uid, gid); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index 836f29480be6..e3a6e2404d11 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -47,7 +47,8 @@ static inline int gfs2_quota_lock_check(struct gfs2_inode *ip, int ret; ap->allowed = UINT_MAX; /* Assume we are permitted a whole lot */ - if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + if (capable(CAP_SYS_RESOURCE) || + sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) return 0; ret = gfs2_quota_lock(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE); if (ret) -- 2.25.1
