[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
atomic_t variables are currently used to implement reference counters with the following properties: - counter is initialized to 1 using atomic_set() - a resource is freed upon counter reaching zero - once counter reaches zero, its further increments aren't allowed - counter schema uses basic atomic operations (set, inc, inc_not_zero, dec_and_test, etc.) Such atomic variables should be converted to a newly provided refcount_t type and API that prevents accidental counter overflows and underflows. This is important since overflows and underflows can lead to use-after-free situation and be exploitable. The variable xlog_ticket.t_ref is used as pure reference counter. Convert it to refcount_t and fix up the operations. **Important note for maintainers: Some functions from refcount_t API defined in lib/refcount.c have different memory ordering guarantees than their atomic counterparts. The full comparison can be seen in https://lkml.org/lkml/2017/11/15/57 and it is hopefully soon in state to be merged to the documentation tree. Normally the differences should not matter since refcount_t provides enough guarantees to satisfy the refcounting use cases, but in some rare cases it might matter. Please double check that you don't have some undocumented memory guarantees for this variable usage. For the xlog_ticket.t_ref it might make a difference in following places: - xfs_log_ticket_put(): decrement in refcount_dec_and_test() only provides RELEASE ordering and control dependency on success vs. fully ordered atomic counterpart Suggested-by: Kees CookReviewed-by: David Windsor Reviewed-by: Hans Liljestrand Signed-off-by: Elena Reshetova --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index a503af9..6e1528f 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3600,8 +3600,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3609,8 +3609,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3732,7 +3732,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 1299759..7482145 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
atomic_t variables are currently used to implement reference counters with the following properties: - counter is initialized to 1 using atomic_set() - a resource is freed upon counter reaching zero - once counter reaches zero, its further increments aren't allowed - counter schema uses basic atomic operations (set, inc, inc_not_zero, dec_and_test, etc.) Such atomic variables should be converted to a newly provided refcount_t type and API that prevents accidental counter overflows and underflows. This is important since overflows and underflows can lead to use-after-free situation and be exploitable. The variable xlog_ticket.t_ref is used as pure reference counter. Convert it to refcount_t and fix up the operations. **Important note for maintainers: Some functions from refcount_t API defined in lib/refcount.c have different memory ordering guarantees than their atomic counterparts. The full comparison can be seen in https://lkml.org/lkml/2017/11/15/57 and it is hopefully soon in state to be merged to the documentation tree. Normally the differences should not matter since refcount_t provides enough guarantees to satisfy the refcounting use cases, but in some rare cases it might matter. Please double check that you don't have some undocumented memory guarantees for this variable usage. For the xlog_ticket.t_ref it might make a difference in following places: - xfs_log_ticket_put(): decrement in refcount_dec_and_test() only provides RELEASE ordering and control dependency on success vs. fully ordered atomic counterpart Suggested-by: Kees Cook Reviewed-by: David Windsor Reviewed-by: Hans Liljestrand Signed-off-by: Elena Reshetova --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index a503af9..6e1528f 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3600,8 +3600,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3609,8 +3609,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3732,7 +3732,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 1299759..7482145 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
atomic_t variables are currently used to implement reference counters with the following properties: - counter is initialized to 1 using atomic_set() - a resource is freed upon counter reaching zero - once counter reaches zero, its further increments aren't allowed - counter schema uses basic atomic operations (set, inc, inc_not_zero, dec_and_test, etc.) Such atomic variables should be converted to a newly provided refcount_t type and API that prevents accidental counter overflows and underflows. This is important since overflows and underflows can lead to use-after-free situation and be exploitable. The variable xlog_ticket.t_ref is used as pure reference counter. Convert it to refcount_t and fix up the operations. Suggested-by: Kees CookReviewed-by: David Windsor Reviewed-by: Hans Liljestrand Signed-off-by: Elena Reshetova --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index dc95a49..e4578c0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3571,8 +3571,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3580,8 +3580,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3703,7 +3703,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 51bf7b8..29f6e1f 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
atomic_t variables are currently used to implement reference counters with the following properties: - counter is initialized to 1 using atomic_set() - a resource is freed upon counter reaching zero - once counter reaches zero, its further increments aren't allowed - counter schema uses basic atomic operations (set, inc, inc_not_zero, dec_and_test, etc.) Such atomic variables should be converted to a newly provided refcount_t type and API that prevents accidental counter overflows and underflows. This is important since overflows and underflows can lead to use-after-free situation and be exploitable. The variable xlog_ticket.t_ref is used as pure reference counter. Convert it to refcount_t and fix up the operations. Suggested-by: Kees Cook Reviewed-by: David Windsor Reviewed-by: Hans Liljestrand Signed-off-by: Elena Reshetova --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index dc95a49..e4578c0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3571,8 +3571,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3580,8 +3580,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3703,7 +3703,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 51bf7b8..29f6e1f 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
RE: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
> On Wed, Mar 8, 2017 at 7:50 AM, Christoph Hellwigwrote: > >> - ASSERT(atomic_read(>t_ref) > 0); > >> - atomic_inc(>t_ref); > >> + ASSERT(refcount_read(>t_ref) > 0); > >> + refcount_inc(>t_ref); > > > > With strict refcount semantics refcount_inc should check that > > the count is larger than 0, otherwise we'd need to use > > recount_inc_not_zero or whatever you're going to call it. > > > > Is that something the recount code does / could do? > > Yes, refcount_inc() will not increment from 0 and WARNs. It looks like > xfs's ASSERT is also a warn (though with XFS-specific formatting), so > perhaps the ASSERT could be dropped? IIUC, Elena's approach to these > changes was to be conservative about removing the existing checks. I am removing the existing WARNs now where they are 100% not needed, but leaving the ones like this ASSERT, because I am thinking that you might have test cases that are dependent on these ASSERTs with specific formatting and I don't want to break them. If you don't need them, I can remove them also. Best Regards, Elena > > -Kees > > -- > Kees Cook > Pixel Security
RE: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
> On Wed, Mar 8, 2017 at 7:50 AM, Christoph Hellwig wrote: > >> - ASSERT(atomic_read(>t_ref) > 0); > >> - atomic_inc(>t_ref); > >> + ASSERT(refcount_read(>t_ref) > 0); > >> + refcount_inc(>t_ref); > > > > With strict refcount semantics refcount_inc should check that > > the count is larger than 0, otherwise we'd need to use > > recount_inc_not_zero or whatever you're going to call it. > > > > Is that something the recount code does / could do? > > Yes, refcount_inc() will not increment from 0 and WARNs. It looks like > xfs's ASSERT is also a warn (though with XFS-specific formatting), so > perhaps the ASSERT could be dropped? IIUC, Elena's approach to these > changes was to be conservative about removing the existing checks. I am removing the existing WARNs now where they are 100% not needed, but leaving the ones like this ASSERT, because I am thinking that you might have test cases that are dependent on these ASSERTs with specific formatting and I don't want to break them. If you don't need them, I can remove them also. Best Regards, Elena > > -Kees > > -- > Kees Cook > Pixel Security
Re: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
On Wed, Mar 8, 2017 at 7:50 AM, Christoph Hellwigwrote: >> - ASSERT(atomic_read(>t_ref) > 0); >> - atomic_inc(>t_ref); >> + ASSERT(refcount_read(>t_ref) > 0); >> + refcount_inc(>t_ref); > > With strict refcount semantics refcount_inc should check that > the count is larger than 0, otherwise we'd need to use > recount_inc_not_zero or whatever you're going to call it. > > Is that something the recount code does / could do? Yes, refcount_inc() will not increment from 0 and WARNs. It looks like xfs's ASSERT is also a warn (though with XFS-specific formatting), so perhaps the ASSERT could be dropped? IIUC, Elena's approach to these changes was to be conservative about removing the existing checks. -Kees -- Kees Cook Pixel Security
Re: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
On Wed, Mar 8, 2017 at 7:50 AM, Christoph Hellwig wrote: >> - ASSERT(atomic_read(>t_ref) > 0); >> - atomic_inc(>t_ref); >> + ASSERT(refcount_read(>t_ref) > 0); >> + refcount_inc(>t_ref); > > With strict refcount semantics refcount_inc should check that > the count is larger than 0, otherwise we'd need to use > recount_inc_not_zero or whatever you're going to call it. > > Is that something the recount code does / could do? Yes, refcount_inc() will not increment from 0 and WARNs. It looks like xfs's ASSERT is also a warn (though with XFS-specific formatting), so perhaps the ASSERT could be dropped? IIUC, Elena's approach to these changes was to be conservative about removing the existing checks. -Kees -- Kees Cook Pixel Security
Re: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
> - ASSERT(atomic_read(>t_ref) > 0); > - atomic_inc(>t_ref); > + ASSERT(refcount_read(>t_ref) > 0); > + refcount_inc(>t_ref); With strict refcount semantics refcount_inc should check that the count is larger than 0, otherwise we'd need to use recount_inc_not_zero or whatever you're going to call it. Is that something the recount code does / could do?
Re: [PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
> - ASSERT(atomic_read(>t_ref) > 0); > - atomic_inc(>t_ref); > + ASSERT(refcount_read(>t_ref) > 0); > + refcount_inc(>t_ref); With strict refcount semantics refcount_inc should check that the count is larger than 0, otherwise we'd need to use recount_inc_not_zero or whatever you're going to call it. Is that something the recount code does / could do?
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
refcount_t type and corresponding API should be used instead of atomic_t when the variable is used as a reference counter. This allows to avoid accidental refcounter overflows that might lead to use-after-free situations. Signed-off-by: Elena ReshetovaSigned-off-by: Hans Liljestrand Signed-off-by: Kees Cook Signed-off-by: David Windsor --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b1469f0..c127fa0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3500,8 +3500,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3509,8 +3509,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3632,7 +3632,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c2604a5..3fc4aba 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
refcount_t type and corresponding API should be used instead of atomic_t when the variable is used as a reference counter. This allows to avoid accidental refcounter overflows that might lead to use-after-free situations. Signed-off-by: Elena Reshetova Signed-off-by: Hans Liljestrand Signed-off-by: Kees Cook Signed-off-by: David Windsor --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b1469f0..c127fa0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3500,8 +3500,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3509,8 +3509,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3632,7 +3632,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c2604a5..3fc4aba 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -168,7 +168,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
refcount_t type and corresponding API should be used instead of atomic_t when the variable is used as a reference counter. This allows to avoid accidental refcounter overflows that might lead to use-after-free situations. Signed-off-by: Elena ReshetovaSigned-off-by: Hans Liljestrand Signed-off-by: Kees Cook Signed-off-by: David Windsor --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 4 +++- 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b1469f0..c127fa0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3500,8 +3500,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3509,8 +3509,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3632,7 +3632,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c2604a5..279afce 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -18,6 +18,8 @@ #ifndef__XFS_LOG_PRIV_H__ #define __XFS_LOG_PRIV_H__ +#include + struct xfs_buf; struct xlog; struct xlog_ticket; @@ -168,7 +170,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4
[PATCH 3/5] fs, xfs: convert xlog_ticket.t_ref from atomic_t to refcount_t
refcount_t type and corresponding API should be used instead of atomic_t when the variable is used as a reference counter. This allows to avoid accidental refcounter overflows that might lead to use-after-free situations. Signed-off-by: Elena Reshetova Signed-off-by: Hans Liljestrand Signed-off-by: Kees Cook Signed-off-by: David Windsor --- fs/xfs/xfs_log.c | 10 +- fs/xfs/xfs_log_priv.h | 4 +++- 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b1469f0..c127fa0 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3500,8 +3500,8 @@ void xfs_log_ticket_put( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - if (atomic_dec_and_test(>t_ref)) + ASSERT(refcount_read(>t_ref) > 0); + if (refcount_dec_and_test(>t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); } @@ -3509,8 +3509,8 @@ xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { - ASSERT(atomic_read(>t_ref) > 0); - atomic_inc(>t_ref); + ASSERT(refcount_read(>t_ref) > 0); + refcount_inc(>t_ref); return ticket; } @@ -3632,7 +3632,7 @@ xlog_ticket_alloc( unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); - atomic_set(>t_ref, 1); + refcount_set(>t_ref, 1); tic->t_task = current; INIT_LIST_HEAD(>t_queue); tic->t_unit_res = unit_res; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c2604a5..279afce 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -18,6 +18,8 @@ #ifndef__XFS_LOG_PRIV_H__ #define __XFS_LOG_PRIV_H__ +#include + struct xfs_buf; struct xlog; struct xlog_ticket; @@ -168,7 +170,7 @@ typedef struct xlog_ticket { struct list_head t_queue; /* reserve/write queue */ struct task_struct *t_task; /* task that owns this ticket */ xlog_tid_t t_tid;/* transaction identifier : 4 */ - atomic_t t_ref;/* ticket reference count : 4 */ + refcount_t t_ref;/* ticket reference count : 4 */ intt_curr_res; /* current reservation in bytes : 4 */ intt_unit_res; /* unit reservation in bytes: 4 */ char t_ocnt; /* original count : 1 */ -- 2.7.4