iommu/vt-d: avoid invalid memory access via node_online(NUMA_NO_NODE)
Subsystems (3)
PATCH 1/1 iommu/vt-d: avoid invalid memory access via node_online(NUMA_NO_NODE)
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)
after converting the type of the first argument (@nr, bit number) of arch_test_bit() from long to unsigned long[0].
Under certain conditions (for example, when ACPI NUMA is disabled via command line), pxm_to_node() can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online() eventually descends to test_bit() without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.
For now, add an explicit check for @node being not %NUMA_NO_NODE before calling test_bit(). The actual logics didn't change here at all.
--- drivers/iommu/intel/dmar.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iommu/intel/dmar.c +1 -1
@@ -494,7 +494,7 @@ static int dmar_parse_one_rhsa(struct acpi_dmar_header *header, void *arg)if (drhd->reg_base_addr == rhsa->base_address) {int node =pxm_to_node (rhsa->proximity_domain); -if (!node_online (node)) +if (node !=NUMA_NO_NODE && !node_online (node)) node =NUMA_NO_NODE ; drhd->iommu->node = node;return 0 ;
Signature · 1 line
2.36.1
From: Alexander Lobakin <[email protected]>
Date: Tue, 12 Jul 2022 17:38:36 +0200
Quoted 23 lines
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)after converting the type of the first argument (@nr, bit number) of
arch_test_bit()fromlongtounsigned long[0].Under certain conditions (for example, when ACPI NUMA is disabled via command line),
pxm_to_node()can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online()eventually descends totest_bit()without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.
Quoted 3 lines
For now, add an explicit check for @node being not %NUMA_NO_NODE before calling
test_bit(). The actual logics didn't change here at all.
Bah, forgot to insert the link here. Hope not worth resending ._.
[0] https://github.com/norov/linux/commit/0e862838f290147ea9c16db852d8d494b552d38d
Quoted 24 lines
Fixes ee34b32d8c29 ("dmar: support for parsing Remapping Hardware Static Affinity structure")Cc [email protected] # 2.6.33+Reported-by kernel test robot <[email protected]>Signed-off-by Alexander Lobakin <[email protected]>--- drivers/iommu/intel/dmar.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/iommu/intel/dmar.c +1
-1@@ -494,7 +494,7 @@ static int dmar_parse_one_rhsa(struct acpi_dmar_header *header, void *arg)if (drhd->reg_base_addr == rhsa->base_address) {int node =pxm_to_node (rhsa->proximity_domain); -if (!node_online (node)) +if (node !=NUMA_NO_NODE && !node_online (node)) node =NUMA_NO_NODE ; drhd->iommu->node = node;return 0 ;Signature · 1 line
2.36.1
Thanks,
Olek
Quoted 27 lines · Alexander Lobakin
From: Alexander Lobakin <[email protected]>
Date: Tue, 12 Jul 2022 17:38:36 +0200Quoted 23 lines
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)after converting the type of the first argument (@nr, bit number) of
arch_test_bit()fromlongtounsigned long[0].Under certain conditions (for example, when ACPI NUMA is disabled via command line),
pxm_to_node()can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online()eventually descends totest_bit()without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.
Quoted 7 lines
Quoted 3 lines
For now, add an explicit check for @node being not %NUMA_NO_NODE before calling
test_bit(). The actual logics didn't change here at all.Bah, forgot to insert the link here. Hope not worth resending ._.
[0] https://github.com/norov/linux/commit/0e862838f290147ea9c16db852d8d494b552d38d
I'll add this link and apply the patch to the bitmap-for-next, after some testing.
Thanks,
Yury
Quoted 33 lines · Yury Norov
Quoted 31 lines · Alexander Lobakin
From: Alexander Lobakin <[email protected]>
Date: Tue, 12 Jul 2022 17:38:36 +0200Quoted 26 lines
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)after converting the type of the first argument (@nr, bit number) of
arch_test_bit()fromlongtounsigned long[0].Under certain conditions (for example, when ACPI NUMA is disabled via command line),
pxm_to_node()can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online()eventually descends totest_bit()without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.For now, add an explicit check for @node being not %NUMA_NO_NODE before calling
test_bit(). The actual logics didn't change here at all.
Quoted 6 lines
Quoted 3 lines
Bah, forgot to insert the link here. Hope not worth resending ._.
[0] https://github.com/norov/linux/commit/0e862838f290147ea9c16db852d8d494b552d38d
I'll add this link and apply the patch to the bitmap-for-next, after some testing.
Thank you!
Best regards,
baolu
Quoted 23 lines · Alexander Lobakin
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)after converting the type of the first argument (@nr, bit number) of
arch_test_bit()fromlongtounsigned long[0].Under certain conditions (for example, when ACPI NUMA is disabled via command line),
pxm_to_node()can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online()eventually descends totest_bit()without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.
Quoted 3 lines
For now, add an explicit check for @node being not %NUMA_NO_NODE before calling
test_bit(). The actual logics didn't change here at all.
Yes, and bitops performance is critical, so it's caller's responsibility to supply correct bit number.
Quoted 24 lines
Fixes ee34b32d8c29 ("dmar: support for parsing Remapping Hardware Static Affinity structure")Cc [email protected] # 2.6.33+Reported-by kernel test robot <[email protected]>Signed-off-by Alexander Lobakin <[email protected]>--- drivers/iommu/intel/dmar.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/iommu/intel/dmar.c +1
-1@@ -494,7 +494,7 @@ static int dmar_parse_one_rhsa(struct acpi_dmar_header *header, void *arg)if (drhd->reg_base_addr == rhsa->base_address) {int node =pxm_to_node (rhsa->proximity_domain); -if (!node_online (node)) +if (node !=NUMA_NO_NODE && !node_online (node)) node =NUMA_NO_NODE ; drhd->iommu->node = node;return 0 ;Signature · 1 line
2.36.1
Signature · 2 lines
With Best Regards, Andy Shevchenko
Yury Norov
AppliedQuoted 23 lines · Alexander Lobakin
KASAN reports:
[ 4.668325][ T0] BUG: KASAN: wild-memory-access in dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497) [ 4.676149][ T0] Read of size 8 at addr 1fffffff85115558 by task swapper/0/0 [ 4.683454][ T0] [ 4.685638][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.19.0-rc3-00004-g0e862838f290 #1 [ 4.694331][ T0] Hardware name: Supermicro SYS-5018D-FN4T/X10SDV-8C-TLN4F, BIOS 1.1 03/02/2016 [ 4.703196][ T0] Call Trace: [ 4.706334][ T0] <TASK> [ 4.709133][ T0] ? dmar_parse_one_rhsa (arch/x86/include/asm/bitops.h:214 arch/x86/include/asm/bitops.h:226 include/asm-generic/bitops/instrumented-non-atomic.h:142 include/linux/nodemask.h:415 drivers/iommu/intel/dmar.c:497)after converting the type of the first argument (@nr, bit number) of
arch_test_bit()fromlongtounsigned long[0].Under certain conditions (for example, when ACPI NUMA is disabled via command line),
pxm_to_node()can return %NUMA_NO_NODE (-1). It is valid 'magic' number of NUMA node, but not valid bit number to use in bitops.node_online()eventually descends totest_bit()without checking for the input, assuming it's on caller side (which might be good for perf-critical tasks). There, -1 becomes %ULONG_MAX which leads to an insane array index when calculating bit position in memory.
Quoted 8 lines
For now, add an explicit check for @node being not %NUMA_NO_NODE before calling
test_bit(). The actual logics didn't change here at all.Fixes ee34b32d8c29 ("dmar: support for parsing Remapping Hardware Static Affinity structure")Cc [email protected] # 2.6.33+Reported-by kernel test robot <[email protected]>Signed-off-by Alexander Lobakin <[email protected]>
Applied, thanks!
Thanks,
Yury
Quoted 19 lines
--- drivers/iommu/intel/dmar.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/iommu/intel/dmar.c +1
-1@@ -494,7 +494,7 @@ static int dmar_parse_one_rhsa(struct acpi_dmar_header *header, void *arg)if (drhd->reg_base_addr == rhsa->base_address) {int node =pxm_to_node (rhsa->proximity_domain); -if (!node_online (node)) +if (node !=NUMA_NO_NODE && !node_online (node)) node =NUMA_NO_NODE ; drhd->iommu->node = node;return 0 ;Signature · 1 line
2.36.1