Skip to contentSkip to conversation

SCSI: mvsas: Fix NULL pointer dereference in mvs_slot_task_free

PATCH In review Dāvis Mosāns 3 participants 9 messages Aug 21, 2015 – Oct 21, 2015 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 SCSI: mvsas: Fix NULL pointer dereference in mvs_slot_task_free

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +2 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info *mvi, u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct sas_task *task,
 			  struct mvs_slot_info *slot, u32 slot_idx)
 {
+	if (!slot)
+		return;
 	if (!slot->task)
 		return;
 	if (!sas_protocol_ata(task->task_proto))
Signature · 6 lines
2.5.0

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#1 lore raw

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 26 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +2 -1
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info *mvi, u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)
                return;
        if (!sas_protocol_ata(task->task_proto))
--
Quoted 2 lines

2.5.0

ping? do I need to CC someone else?
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Reviewed-by on v1 1/1
#2 lore raw
In reply to Dāvis Mosāns
Quoted 8 lines · Dāvis Mosāns

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

Cheers,
Tomas

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#3 lore raw
In reply to Dāvis Mosāns

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 26 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +2 -1
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info *mvi, u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)
                return;
        if (!sas_protocol_ata(task->task_proto))
--
Quoted 2 lines

2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

Reviewed-by
#4 lore raw
Quoted 28 lines · Dāvis Mosāns

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 27 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays
NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++

 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +0 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info *mvi,
 u32 rx_desc)

 static void mvs_slot_task_free(struct mvs_info *mvi, struct sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx) {
+       if (!slot)
+               return;
        if (!slot->task)

Quoted 8 lines
Quoted 5 lines

                return;
        if (!sas_protocol_ata(task->task_proto))
--
2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

If it saved you from several crashes, it probably should be tagged for stable, shouldn't it?

Quoted 5 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

#5 lore raw

2015-10-21 10:33 GMT+03:00 Johannes Thumshirn <[email protected]>:

Quoted 38 lines · Dāvis Mosāns
Quoted 28 lines

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 27 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays
NULL but it's later used in mvs_abort_task as slot which is passed to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +2 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info *mvi,
 u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct
 sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)
Quoted 8 lines
Quoted 5 lines
                return;
        if (!sas_protocol_ata(task->task_proto))

--
2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

Quoted 6 lines

If it saved you from several crashes, it probably should be tagged for stable, shouldn't it?

I don't really know how that works... this is my first patch so I'm not really concerned about in which version it gets in as long as it does. I've been compiling kernel with this patch for these months so for me it doesn't really make any difference.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#6 lore raw
Quoted 50 lines · Dāvis Mosāns

2015-10-21 10:33 GMT+03:00 Johannes Thumshirn <[email protected]>:

Quoted 41 lines · Dāvis Mosāns
Quoted 31 lines

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 30 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays
NULL but it's later used in mvs_abort_task as slot which is passed
to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++

 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +0 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info
 *mvi,
 u32 rx_desc)

 static void mvs_slot_task_free(struct mvs_info *mvi, struct sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)

Quoted 8 lines
Quoted 5 lines

                return;
        if (!sas_protocol_ata(task->task_proto))
--
2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

Quoted 7 lines

If it saved you from several crashes, it probably should be tagged for
stable, shouldn't it?

Quoted 6 lines

I don't really know how that works... this is my first patch so I'm not really concerned about in which version it gets in as long as it does.
I've been compiling kernel with this patch for these months so for me it
doesn't really make any difference.

You can add
Cc: [email protected]
somewhere around your Signed-off-by

Documentation/stable_kernel_rules.txt has all the process documentation.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#7 lore raw

2015-10-21 16:47 GMT+03:00 Johannes Thumshirn <[email protected]>:

Quoted 58 lines · Dāvis Mosāns
Quoted 50 lines

2015-10-21 10:33 GMT+03:00 Johannes Thumshirn <[email protected]>:

Quoted 41 lines · Dāvis Mosāns
Quoted 31 lines

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 30 lines

When pci_pool_alloc fails in mvs_task_prep then task->lldd_task stays
NULL but it's later used in mvs_abort_task as slot which is passed
to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=101891

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +2 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct mvs_info
 *mvi,
 u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct
 sas_task *task,
                          struct mvs_slot_info *slot, u32
 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)
Quoted 8 lines
Quoted 5 lines
                return;
        if (!sas_protocol_ata(task->task_proto))

--
2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

Quoted 7 lines

If it saved you from several crashes, it probably should be tagged for
stable, shouldn't it?

Quoted 6 lines

I don't really know how that works... this is my first patch so I'm not really concerned about in which version it gets in as long as it does.
I've been compiling kernel with this patch for these months so for me it
doesn't really make any difference.

Quoted 6 lines

You can add
Cc: [email protected]
somewhere around your Signed-off-by

Documentation/stable_kernel_rules.txt has all the process documentation.

Should I add it together with review tags too and resend patch or reply to this thread with it edited or just leave it like it is now and whoever will see it will add it himself?

also for stable requirements this line is a bit confusing "It or an equivalent fix must already exist in Linus' tree (upstream)."

but then later seems it's not requirement for Option 1

#8 lore raw
Quoted 74 lines · Dāvis Mosāns

2015-10-21 16:47 GMT+03:00 Johannes Thumshirn <[email protected]>:

Quoted 73 lines · Dāvis Mosāns
Quoted 58 lines

2015-10-21 10:33 GMT+03:00 Johannes Thumshirn <[email protected]

Quoted 48 lines

:

Quoted 37 lines · Dāvis Mosāns

2015-08-21 7:29 GMT+03:00 Dāvis Mosāns <[email protected]>:

Quoted 36 lines

When pci_pool_alloc fails in mvs_task_prep then task-

Quoted 1 line

lldd_task

stays
NULL but it's later used in mvs_abort_task as slot which is passed
to mvs_slot_task_free causing NULL pointer dereference.

Just return from mvs_slot_task_free when passed with NULL slot.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=10189
1

---
 drivers/scsi/mvsas/mv_sas.c | 2 ++

 1 file changed, 2 insertions(+)
drivers/scsi/mvsas/mv_sas.c +0 -0
index 454536c..9c78074 100644
@@ -887,6 +887,8 @@ static void mvs_slot_free(struct

mvs_info
*mvi,
u32 rx_desc)
 static void mvs_slot_task_free(struct mvs_info *mvi, struct
sas_task *task,
                          struct mvs_slot_info *slot, u32 slot_idx)
 {
+       if (!slot)
+               return;
        if (!slot->task)
                return;

Quoted 8 lines
Quoted 4 lines

        if (!sas_protocol_ata(task->task_proto))
--
2.5.0

Can this get merged?
So far since august it have saved me from several kernel crashes.

Quoted 8 lines

If it saved you from several crashes, it probably should be tagged
for
stable, shouldn't it?

Quoted 9 lines

I don't really know how that works... this is my first patch so I'm
not really concerned about in which version it gets in as long as it
does.
I've been compiling kernel with this patch for these months so for me
it
doesn't really make any difference.

You can add
Cc: [email protected]
somewhere around your Signed-off-by

Quoted 8 lines
Quoted 2 lines

Documentation/stable_kernel_rules.txt has all the process documentation.

Should I add it together with review tags too and resend patch or reply to this
thread with it edited or just leave it like it is now and whoever will see it
will add it himself?

good question, but I think James can help here.

Quoted 7 lines

also for stable requirements this line is a bit confusing "It or an equivalent fix must already exist in Linus' tree (upstream)."

but then later seems it's not requirement for Option 1

yes, if you tag it with the Cc, it will get to stable review _after_ it is applied to Linus' tree