fix sbitmap initialization and null_blk shared tagset behavior
Subsystems (2)
v3 cover fix sbitmap initialization and null_blk shared tagset behavior
Hi,
This patchset fixes two subtle issues discovered while unit testing nr_hw_queue update code using null_blk driver.
The first patch in the series, fixes an issue in the sbitmap initialization code, where sb->alloc_hint is not explicitly set to NULL when the sbitmap depth is zero. This can lead to a kernel crash in sbitmap_free(), which unconditionally calls free_percpu() on sb->alloc_hint — even if it was never allocated. The crash is caused by dereferencing an invalid pointer or stale garbage value.
The second patch in the series, prevents runtime updates to submit_queues or poll_queues when using a shared tagset. Currently, such updates lead to the allocation of new hardware queues (hctx) that are never mapped to any software queues (ctx), rendering them unusable for I/O. This patch rejects these changes and ensures more consistent behavior. Interestingly, this unnecessary queue update path helped uncover the issue fixed in first patch.
As usual, review and feedback are most welcome!
Changes from v2:
- Updated the second patch to prevent the user from modifying submit or poll queues when tagset is shared (Damien Le Moal, Yu Kuai)
Changes from v1:
- The set->driver_data field should be initialized separately for the shared tagset to ensure it is correctly set for both shared and non-shared tagset cases. (Damien Le Moal)
Nilay Shroff (2):
lib/sbitmap: fix kernel crash observed when sbitmap depth is zero
null_blk: prevent submit and poll queues update for shared tagset
drivers/block/null_blk/main.c | 32 ++++++++++++++++++++++---------- lib/sbitmap.c | 1 + 2 files changed, 23 insertions(+), 10 deletions(-)
Signature · 1 line
2.50.1
PATCH v3 1/2 lib/sbitmap: fix kernel crash observed when sbitmap depth is zero
We observed a kernel crash when the I/O scheduler allocates an sbitmap for a hardware queue (hctx) that has no associated software queues (ctx), and later attempts to free it. When no software queues are mapped to a hardware queue, the sbitmap is initialized with a depth of zero. In such cases, the sbitmap_init_node() function should set sb->alloc_hint to NULL. However, if this is not done, sb->alloc_hint may contain garbage, and calling sbitmap_free() will pass this invalid pointer to free_percpu(), resulting in a kernel crash.
Example crash trace:
==================================================================
Kernel attempted to read user page (28) - exploit attempt? (uid: 0)
BUG: Kernel NULL pointer dereference on read at 0x00000028
Faulting instruction address: 0xc000000000708f88
Oops: Kernel access of bad area, sig: 11 [#1]
LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
[...]
CPU: 5 UID: 0 PID: 5491 Comm: mk_nullb_shared Kdump: loaded Tainted: G B 6.16.0-rc5+ #294 VOLUNTARY
Tainted: [B]=BAD_PAGE
Hardware name: IBM,9043-MRX POWER10 (architected) 0x800200 0xf000006 of:IBM,FW1060.00 (NM1060_028) hv:phyp pSeries
[...]
NIP [c000000000708f88] free_percpu+0x144/0xba8
LR [c000000000708f84] free_percpu+0x140/0xba8
Call Trace:
free_percpu+0x140/0xba8 (unreliable)
kyber_exit_hctx+0x94/0x124
blk_mq_exit_sched+0xe4/0x214
elevator_exit+0xa8/0xf4
elevator_switch+0x3b8/0x5d8
elv_update_nr_hw_queues+0x14c/0x300
blk_mq_update_nr_hw_queues+0x5cc/0x670
nullb_update_nr_hw_queues+0x118/0x1f8 [null_blk]
nullb_device_submit_queues_store+0xac/0x170 [null_blk]
configfs_write_iter+0x1dc/0x2d0
vfs_write+0x5b0/0x77c
ksys_write+0xa0/0x180
system_call_exception+0x1b0/0x4f0
system_call_vectored_common+0x15c/0x2ecIf the sbitmap depth is zero, sb->alloc_hint memory is NOT allocated, but the pointer is not explicitly set to NULL. Later, during sbitmap_free(), the kernel attempts to free sb->alloc_hint, which is a per cpu pointer variable, regardless of whether it was valid, leading to a crash.
This patch ensures that sb->alloc_hint is explicitly set to NULL in sbitmap_init_node() when the requested depth is zero. This prevents free_percpu() from freeing sb->alloc_hint and thus avoids the observed crash.
--- lib/sbitmap.c | 1 + 1 file changed, 1 insertion(+)
lib/sbitmap.c +1 -0
@@ -119,6 +119,7 @@ int sbitmap_init_node(struct sbitmap *sb, unsigned int depth, int shift,if (depth ==0 ) { sb->map =NULL ; + sb->alloc_hint =NULL ;return 0 ; }
Signature · 1 line
2.50.1
Re: [PATCHv3 1/2] lib/sbitmap: fix kernel crash observed when sbitmap depth is zero
Quoted 49 lines · Nilay Shroff
We observed a kernel crash when the I/O scheduler allocates an sbitmap for a hardware queue (hctx) that has no associated software queues (ctx), and later attempts to free it. When no software queues are mapped to a hardware queue, the sbitmap is initialized with a depth of zero. In such cases, the
sbitmap_init_node()function should setsb->alloc_hintto NULL. However, if this is not done,sb->alloc_hintmay contain garbage, and callingsbitmap_free()will pass this invalid pointer tofree_percpu(), resulting in a kernel crash.Example crash trace:
==================================================================
Kernel attempted to read user page (28) - exploit attempt? (uid: 0)BUG: Kernel NULL pointer dereference on read at 0x00000028 Faulting instruction address: 0xc000000000708f88 Oops: Kernel access of bad area, sig: 11 [#1] LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries [...] CPU: 5 UID: 0 PID: 5491 Comm: mk_nullb_shared Kdump: loaded Tainted: G B 6.16.0-rc5+ #294 VOLUNTARY Tainted: [B]=BAD_PAGE Hardware name: IBM,9043-MRX POWER10 (architected) 0x800200 0xf000006 of:IBM,FW1060.00 (NM1060_028) hv:phyp pSeries [...] NIP [c000000000708f88] free_percpu+0x144/0xba8 LR [c000000000708f84] free_percpu+0x140/0xba8 Call Trace: free_percpu+0x140/0xba8 (unreliable) kyber_exit_hctx+0x94/0x124 blk_mq_exit_sched+0xe4/0x214 elevator_exit+0xa8/0xf4 elevator_switch+0x3b8/0x5d8 elv_update_nr_hw_queues+0x14c/0x300 blk_mq_update_nr_hw_queues+0x5cc/0x670 nullb_update_nr_hw_queues+0x118/0x1f8 [null_blk] nullb_device_submit_queues_store+0xac/0x170 [null_blk] configfs_write_iter+0x1dc/0x2d0 vfs_write+0x5b0/0x77c ksys_write+0xa0/0x180 system_call_exception+0x1b0/0x4f0 system_call_vectored_common+0x15c/0x2ecIf the sbitmap depth is zero,
sb->alloc_hintmemory is NOT allocated, but the pointer is not explicitly set to NULL. Later, duringsbitmap_free(), the kernel attempts to freesb->alloc_hint, which is a per cpu pointer variable, regardless of whether it was valid, leading to a crash.This patch ensures that
sb->alloc_hintis explicitly set to NULL insbitmap_init_node()when the requested depth is zero. This preventsfree_percpu()from freeingsb->alloc_hintand thus avoids the observed crash.
Quoted 7 lines
Reviewed-by Damien Le Moal <[email protected]>Reviewed-by Hannes Reinecke <[email protected]>Signed-off-by Nilay Shroff <[email protected]>--- lib/sbitmap.c | 1 + 1 file changed, 1 insertion(+)
Thanks
Quoted 13 lines
lib/sbitmap.c +1
-0@@ -119,6 +119,7 @@ int sbitmap_init_node(struct sbitmap *sb, unsigned int depth, int shift,if (depth ==0 ) { sb->map =NULL ; + sb->alloc_hint =NULL ;return 0 ; }
PATCH v3 2/2 null_blk: prevent submit and poll queues update for shared tagset
In reply to Nilay ShroffWhen a user updates the number of submit or poll queues on a null_blk device, the block layer creates new hardware queues (hctxs). However, if the device is using a shared tagset, null_blk does not map any software queues (ctx) to the newly created hctx (via null_map_queues()), resulting in those hardware queues being left unused for I/O. This behavior is misleading, as the user may expect the new queues to be functional, even though they are effectively ignored. To avoid this confusion and potential misconfiguration:
- Reject runtime updates to submit_queues or poll_queues via sysfs when the device uses a shared tagset by returning -EINVAL.
- During configuration validation (prior to powering on the device), reset submit_queues and poll_queues to the module parameters (g_submit_queues and g_poll_queues) if the shared tagset is enabled.
This ensures consistent behavior and avoids creating unused hardware queues (hctxs) due to ineffective runtime queue updates.
--- drivers/block/null_blk/main.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-)
drivers/block/null_blk/main.c +22 -10
@@ -388,6 +388,12 @@ static int nullb_update_nr_hw_queues(struct nullb_device *dev,if (!submit_queues)return -EINVAL ; +/* +* Cannot update queues with shared tagset. +*/ +if (dev->shared_tags) +return -EINVAL ; +/* * Make sure that null_init_hctx() does not access nullb->queues[] past * the end of that array.
@@ -1884,18 +1890,24 @@ static int null_validate_conf(struct nullb_device *dev) dev->queue_mode =NULL_Q_MQ ; } -if (dev->use_per_node_hctx) { -if (dev->submit_queues != nr_online_nodes) - dev->submit_queues = nr_online_nodes; - }else if (dev->submit_queues > nr_cpu_ids) - dev->submit_queues = nr_cpu_ids; -else if (dev->submit_queues ==0 ) - dev->submit_queues =1 ; - dev->prev_submit_queues = dev->submit_queues; - -if (dev->poll_queues > g_poll_queues) +if (dev->shared_tags) { + dev->submit_queues = g_submit_queues; dev->poll_queues = g_poll_queues; + }else { +if (dev->use_per_node_hctx) { +if (dev->submit_queues != nr_online_nodes) + dev->submit_queues = nr_online_nodes; + }else if (dev->submit_queues > nr_cpu_ids) + dev->submit_queues = nr_cpu_ids; +else if (dev->submit_queues ==0 ) + dev->submit_queues =1 ; + +if (dev->poll_queues > g_poll_queues) + dev->poll_queues = g_poll_queues; + } + dev->prev_submit_queues = dev->submit_queues; dev->prev_poll_queues = dev->poll_queues; + dev->irqmode = min_t(unsigned int, dev->irqmode, NULL_IRQ_TIMER);/* Do memory allocation, so set blocking */
Signature · 1 line
2.50.1
Re: [PATCHv3 2/2] null_blk: prevent submit and poll queues update for shared tagset
Quoted 14 lines · Nilay Shroff
When a user updates the number of submit or poll queues on a
null_blkdevice, the block layer creates new hardware queues (hctxs). However, if the device is using a shared tagset,null_blkdoes not map any software queues (ctx) to the newly created hctx (vianull_map_queues()), resulting in those hardware queues being left unused for I/O. This behavior is misleading, as the user may expect the new queues to be functional, even though they are effectively ignored. To avoid this confusion and potential misconfiguration:
- Reject runtime updates tosubmit_queuesorpoll_queuesvia sysfs when the device uses a shared tagset by returning-EINVAL.
- During configuration validation (prior to powering on the device), resetsubmit_queuesandpoll_queuesto the module parameters (g_submit_queuesandg_poll_queues) if the shared tagset is enabled.
Quoted 8 lines
This ensures consistent behavior and avoids creating unused hardware queues (hctxs) due to ineffective runtime queue updates.
Signed-off-by Nilay Shroff <[email protected]>--- drivers/block/null_blk/main.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-)
Cheers,
Hannes
Signature · 4 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich
Re: [PATCHv3 2/2] null_blk: prevent submit and poll queues update for shared tagset
In reply to Nilay ShroffQuoted 14 lines · Nilay Shroff
When a user updates the number of submit or poll queues on a
null_blkdevice, the block layer creates new hardware queues (hctxs). However, if the device is using a shared tagset,null_blkdoes not map any software queues (ctx) to the newly created hctx (vianull_map_queues()), resulting in those hardware queues being left unused for I/O. This behavior is misleading, as the user may expect the new queues to be functional, even though they are effectively ignored. To avoid this confusion and potential misconfiguration:
- Reject runtime updates tosubmit_queuesorpoll_queuesvia sysfs when the device uses a shared tagset by returning-EINVAL.
- During configuration validation (prior to powering on the device), resetsubmit_queuesandpoll_queuesto the module parameters (g_submit_queuesandg_poll_queues) if the shared tagset is enabled.
Quoted 4 lines
This ensures consistent behavior and avoids creating unused hardware queues (hctxs) due to ineffective runtime queue updates.
Signed-off-by Nilay Shroff <[email protected]>
Signature · 2 lines
Damien Le Moal Western Digital Research
Re: [PATCHv3 2/2] null_blk: prevent submit and poll queues update for shared tagset
In reply to Nilay ShroffQuoted 14 lines · Nilay Shroff
When a user updates the number of submit or poll queues on a
null_blkdevice, the block layer creates new hardware queues (hctxs). However, if the device is using a shared tagset,null_blkdoes not map any software queues (ctx) to the newly created hctx (vianull_map_queues()), resulting in those hardware queues being left unused for I/O. This behavior is misleading, as the user may expect the new queues to be functional, even though they are effectively ignored. To avoid this confusion and potential misconfiguration:
- Reject runtime updates tosubmit_queuesorpoll_queuesvia sysfs when the device uses a shared tagset by returning-EINVAL.
- During configuration validation (prior to powering on the device), resetsubmit_queuesandpoll_queuesto the module parameters (g_submit_queuesandg_poll_queues) if the shared tagset is enabled.
Quoted 8 lines
This ensures consistent behavior and avoids creating unused hardware queues (hctxs) due to ineffective runtime queue updates.
Signed-off-by Nilay Shroff <[email protected]>--- drivers/block/null_blk/main.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-)
Thanks
Quoted 53 lines
drivers/block/null_blk/main.c +22
-10@@ -388,6 +388,12 @@ static int nullb_update_nr_hw_queues(struct nullb_device *dev,if (!submit_queues)return -EINVAL ; +/* +* Cannot update queues with shared tagset. +*/ +if (dev->shared_tags) +return -EINVAL ; +/* * Make sure that null_init_hctx() does not access nullb->queues[] past * the end of that array. @@ -1884,18 +1890,24 @@ static int null_validate_conf(struct nullb_device *dev) dev->queue_mode =NULL_Q_MQ ; } -if (dev->use_per_node_hctx) { -if (dev->submit_queues != nr_online_nodes) - dev->submit_queues = nr_online_nodes; - }else if (dev->submit_queues > nr_cpu_ids) - dev->submit_queues = nr_cpu_ids; -else if (dev->submit_queues ==0 ) - dev->submit_queues =1 ; - dev->prev_submit_queues = dev->submit_queues; - -if (dev->poll_queues > g_poll_queues) +if (dev->shared_tags) { + dev->submit_queues = g_submit_queues; dev->poll_queues = g_poll_queues; + }else { +if (dev->use_per_node_hctx) { +if (dev->submit_queues != nr_online_nodes) + dev->submit_queues = nr_online_nodes; + }else if (dev->submit_queues > nr_cpu_ids) + dev->submit_queues = nr_cpu_ids; +else if (dev->submit_queues ==0 ) + dev->submit_queues =1 ; + +if (dev->poll_queues > g_poll_queues) + dev->poll_queues = g_poll_queues; + } + dev->prev_submit_queues = dev->submit_queues; dev->prev_poll_queues = dev->poll_queues; + dev->irqmode = min_t(unsigned int, dev->irqmode, NULL_IRQ_TIMER);/* Do memory allocation, so set blocking */
Hi Jens,
It seems this patchset may have fallen through the cracks. It still applies cleanly on the latest mainline code. Could you please consider pulling it for the 6.18 cycle?
Thanks,
--Nilay
Quoted 38 lines · Nilay Shroff
Hi,
This patchset fixes two subtle issues discovered while unit testing
nr_hw_queueupdate code usingnull_blkdriver.The first patch in the series, fixes an issue in the sbitmap initialization code, where
sb->alloc_hintis not explicitly set to NULL when the sbitmap depth is zero. This can lead to a kernel crash insbitmap_free(), which unconditionally callsfree_percpu()onsb->alloc_hint— even if it was never allocated. The crash is caused by dereferencing an invalid pointer or stale garbage value.The second patch in the series, prevents runtime updates to
submit_queuesorpoll_queueswhen using a shared tagset. Currently, such updates lead to the allocation of new hardware queues (hctx) that are never mapped to any software queues (ctx), rendering them unusable for I/O. This patch rejects these changes and ensures more consistent behavior. Interestingly, this unnecessary queue update path helped uncover the issue fixed in first patch.As usual, review and feedback are most welcome!
Changes from v2:
- Updated the second patch to prevent the user from modifying submit or poll queues when tagset is shared (Damien Le Moal, Yu Kuai)
Changes from v1:
- Theset->driver_datafield should be initialized separately for the shared tagset to ensure it is correctly set for both shared and non-shared tagset cases. (Damien Le Moal)Nilay Shroff (2):
lib/sbitmap: fix kernel crash observed when sbitmap depth is zero
null_blk: prevent submit and poll queues update for shared tagsetdrivers/block/null_blk/main.c | 32 ++++++++++++++++++++++---------- lib/sbitmap.c | 1 + 2 files changed, 23 insertions(+), 10 deletions(-)