* [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 @ 2022-08-05 9:47 Bruno Goncalves 2022-08-05 12:56 ` Mark Rutland 2022-08-05 13:59 ` Sudeep Holla 0 siblings, 2 replies; 6+ messages in thread From: Bruno Goncalves @ 2022-08-05 9:47 UTC (permalink / raw) To: linux-arm-kernel, LKML; +Cc: CKI Project Hello, Since commit "c1c76700a0d6" we started to hit the following call trace on aarch64 VMs when looking at journalctl log kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 bytes, linear) kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate kernel: ------------[ cut here ]------------ kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 kernel: Modules linked in: kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) kernel: pc : __alloc_pages+0x1a0/0x290 kernel: lr : alloc_page_interleave+0x24/0xbc kernel: sp : ffff80000803bb90 kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 kernel: Call trace: kernel: __alloc_pages+0x1a0/0x290 kernel: alloc_page_interleave+0x24/0xbc kernel: alloc_pages+0x10c/0x16c kernel: kmalloc_order+0x3c/0xc0 kernel: kmalloc_order_trace+0x38/0x130 kernel: __kmalloc+0x2e8/0x350 kernel: detect_cache_attributes+0x9c/0x110 kernel: update_siblings_masks+0x34/0x270 kernel: store_cpu_topology+0x7c/0xc0 kernel: smp_prepare_cpus+0x34/0x108 kernel: kernel_init_freeable+0x108/0x1b8 kernel: kernel_init+0x30/0x150 kernel: ret_from_fork+0x10/0x20 kernel: ---[ end trace 0000000000000000 ]--- kernel: Early cacheinfo failed, ret = -12 kernel: cblist_init_generic: Setting adjustable number of callback queues. kernel: cblist_init_generic: Setting shift to 2 and lim to 1. test logs: https://datawarehouse.cki-project.org/kcidb/tests/4712644 cki issue tracker: https://datawarehouse.cki-project.org/issue/1485 kernel config: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/build%20aarch64/2816138306/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.config kernel build: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/publish%20aarch64/2816138315/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.tar.gz Thanks, Bruno Goncalves ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 2022-08-05 9:47 [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 Bruno Goncalves @ 2022-08-05 12:56 ` Mark Rutland 2022-08-05 13:28 ` Bruno Goncalves 2022-08-05 13:59 ` Sudeep Holla 1 sibling, 1 reply; 6+ messages in thread From: Mark Rutland @ 2022-08-05 12:56 UTC (permalink / raw) To: Bruno Goncalves; +Cc: linux-arm-kernel, LKML, CKI Project, Sudeep Holla On Fri, Aug 05, 2022 at 11:47:57AM +0200, Bruno Goncalves wrote: > Hello, > > Since commit "c1c76700a0d6" we started to hit the following call trace > on aarch64 VMs when looking at journalctl log For everyone else's benefit that's commit: c1c76700a0d6 ("Merge tag 'spdx-6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/spdx") Could you please include the commit title when citing a patch? ... did you try to bisect this? > kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 > bytes, linear) > kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate > kernel: ------------[ cut here ]------------ > kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 IIUC that's: WARN_ON_ONCE_GFP(order >= MAX_ORDER, gfp) ... and IIUC that's fixed by commit: 11969d698f8cda31 ("cacheinfo: Use atomic allocation for percpu cache attributes") Thanks, Mark. > kernel: Modules linked in: > kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 > kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) > kernel: pc : __alloc_pages+0x1a0/0x290 > kernel: lr : alloc_page_interleave+0x24/0xbc > kernel: sp : ffff80000803bb90 > kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 > kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 > kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e > kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 > kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 > kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 > kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 > kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 > kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 > kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 > kernel: Call trace: > kernel: __alloc_pages+0x1a0/0x290 > kernel: alloc_page_interleave+0x24/0xbc > kernel: alloc_pages+0x10c/0x16c > kernel: kmalloc_order+0x3c/0xc0 > kernel: kmalloc_order_trace+0x38/0x130 > kernel: __kmalloc+0x2e8/0x350 > kernel: detect_cache_attributes+0x9c/0x110 > kernel: update_siblings_masks+0x34/0x270 > kernel: store_cpu_topology+0x7c/0xc0 > kernel: smp_prepare_cpus+0x34/0x108 > kernel: kernel_init_freeable+0x108/0x1b8 > kernel: kernel_init+0x30/0x150 > kernel: ret_from_fork+0x10/0x20 > kernel: ---[ end trace 0000000000000000 ]--- > kernel: Early cacheinfo failed, ret = -12 > kernel: cblist_init_generic: Setting adjustable number of callback queues. > kernel: cblist_init_generic: Setting shift to 2 and lim to 1. > > test logs: https://datawarehouse.cki-project.org/kcidb/tests/4712644 > cki issue tracker: https://datawarehouse.cki-project.org/issue/1485 > > kernel config: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/build%20aarch64/2816138306/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.config > kernel build: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/publish%20aarch64/2816138315/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.tar.gz > > Thanks, > Bruno Goncalves > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 2022-08-05 12:56 ` Mark Rutland @ 2022-08-05 13:28 ` Bruno Goncalves 0 siblings, 0 replies; 6+ messages in thread From: Bruno Goncalves @ 2022-08-05 13:28 UTC (permalink / raw) To: Mark Rutland; +Cc: linux-arm-kernel, LKML, CKI Project, Sudeep Holla On Fri, 5 Aug 2022 at 14:57, Mark Rutland <mark.rutland@arm.com> wrote: > > On Fri, Aug 05, 2022 at 11:47:57AM +0200, Bruno Goncalves wrote: > > Hello, > > > > Since commit "c1c76700a0d6" we started to hit the following call trace > > on aarch64 VMs when looking at journalctl log > > For everyone else's benefit that's commit: > > c1c76700a0d6 ("Merge tag 'spdx-6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/spdx") > > Could you please include the commit title when citing a patch? Thanks for the extra information. > > ... did you try to bisect this? No, I didn't bisect it. To add more context here I meant that c1c76700a0d6 ("Merge tag 'spdx-6.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/spdx") is the first commit that we tested that hit the problem. The last commit that we tested without this problem is b44f2fd87919 ("Merge tag 'drm-next-2022-08-03' of git://anongit.freedesktop.org/drm/drm") https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b44f2fd87919b5ae6e1756d4c7ba2cbba22238e1 > > > kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 > > bytes, linear) > > kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate > > kernel: ------------[ cut here ]------------ > > kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 > > IIUC that's: > > WARN_ON_ONCE_GFP(order >= MAX_ORDER, gfp) > > ... and IIUC that's fixed by commit: > > 11969d698f8cda31 ("cacheinfo: Use atomic allocation for percpu cache attributes") I believe this commit is already included in the tree. We've also reproduced it using latest commit "b2a88c212e65" ("Merge tag 'xfs-5.20-merge-6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux") https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b2a88c212e652e94f1e4b635910972ac57ba4e97 Thanks, Bruno > Thanks, > Mark. > > > kernel: Modules linked in: > > kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 > > kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) > > kernel: pc : __alloc_pages+0x1a0/0x290 > > kernel: lr : alloc_page_interleave+0x24/0xbc > > kernel: sp : ffff80000803bb90 > > kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 > > kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 > > kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e > > kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 > > kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 > > kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 > > kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 > > kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 > > kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 > > kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 > > kernel: Call trace: > > kernel: __alloc_pages+0x1a0/0x290 > > kernel: alloc_page_interleave+0x24/0xbc > > kernel: alloc_pages+0x10c/0x16c > > kernel: kmalloc_order+0x3c/0xc0 > > kernel: kmalloc_order_trace+0x38/0x130 > > kernel: __kmalloc+0x2e8/0x350 > > kernel: detect_cache_attributes+0x9c/0x110 > > kernel: update_siblings_masks+0x34/0x270 > > kernel: store_cpu_topology+0x7c/0xc0 > > kernel: smp_prepare_cpus+0x34/0x108 > > kernel: kernel_init_freeable+0x108/0x1b8 > > kernel: kernel_init+0x30/0x150 > > kernel: ret_from_fork+0x10/0x20 > > kernel: ---[ end trace 0000000000000000 ]--- > > kernel: Early cacheinfo failed, ret = -12 > > kernel: cblist_init_generic: Setting adjustable number of callback queues. > > kernel: cblist_init_generic: Setting shift to 2 and lim to 1. > > > > test logs: https://datawarehouse.cki-project.org/kcidb/tests/4712644 > > cki issue tracker: https://datawarehouse.cki-project.org/issue/1485 > > > > kernel config: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/build%20aarch64/2816138306/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.config > > kernel build: https://s3.amazonaws.com/arr-cki-prod-trusted-artifacts/trusted-artifacts/605193135/publish%20aarch64/2816138315/artifacts/kernel-mainline.kernel.org-redhat_605193135_aarch64.tar.gz > > > > Thanks, > > Bruno Goncalves > > > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 2022-08-05 9:47 [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 Bruno Goncalves 2022-08-05 12:56 ` Mark Rutland @ 2022-08-05 13:59 ` Sudeep Holla 2022-08-05 15:08 ` Sudeep Holla 1 sibling, 1 reply; 6+ messages in thread From: Sudeep Holla @ 2022-08-05 13:59 UTC (permalink / raw) To: Bruno Goncalves; +Cc: linux-arm-kernel, Sudeep Holla, LKML, CKI Project On Fri, Aug 05, 2022 at 11:47:57AM +0200, Bruno Goncalves wrote: > Hello, > > Since commit "c1c76700a0d6" we started to hit the following call trace > on aarch64 VMs when looking at journalctl log > > kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 > bytes, linear) > kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate > kernel: ------------[ cut here ]------------ > kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 > kernel: Modules linked in: > kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 > kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) > kernel: pc : __alloc_pages+0x1a0/0x290 > kernel: lr : alloc_page_interleave+0x24/0xbc > kernel: sp : ffff80000803bb90 > kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 > kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 > kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e > kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 > kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 > kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 > kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 > kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 > kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 > kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 > kernel: Call trace: > kernel: __alloc_pages+0x1a0/0x290 > kernel: alloc_page_interleave+0x24/0xbc > kernel: alloc_pages+0x10c/0x16c > kernel: kmalloc_order+0x3c/0xc0 > kernel: kmalloc_order_trace+0x38/0x130 > kernel: __kmalloc+0x2e8/0x350 > kernel: detect_cache_attributes+0x9c/0x110 > kernel: update_siblings_masks+0x34/0x270 > kernel: store_cpu_topology+0x7c/0xc0 > kernel: smp_prepare_cpus+0x34/0x108 > kernel: kernel_init_freeable+0x108/0x1b8 > kernel: kernel_init+0x30/0x150 > kernel: ret_from_fork+0x10/0x20 > kernel: ---[ end trace 0000000000000000 ]--- > kernel: Early cacheinfo failed, ret = -12 > kernel: cblist_init_generic: Setting adjustable number of callback queues. > kernel: cblist_init_generic: Setting shift to 2 and lim to 1. > Thanks for the report. I will see if I can reproduce by forcing the absence of PPTT which is something I haven't tested after I fixed hotplug issues reported with my changes. -- Regards, Sudeep ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 2022-08-05 13:59 ` Sudeep Holla @ 2022-08-05 15:08 ` Sudeep Holla 2022-08-05 18:38 ` Bruno Goncalves 0 siblings, 1 reply; 6+ messages in thread From: Sudeep Holla @ 2022-08-05 15:08 UTC (permalink / raw) To: Bruno Goncalves; +Cc: linux-arm-kernel, Sudeep Holla, LKML, CKI Project On Fri, Aug 05, 2022 at 02:59:55PM +0100, Sudeep Holla wrote: > On Fri, Aug 05, 2022 at 11:47:57AM +0200, Bruno Goncalves wrote: > > Hello, > > > > Since commit "c1c76700a0d6" we started to hit the following call trace > > on aarch64 VMs when looking at journalctl log > > > > kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 > > bytes, linear) > > kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate > > kernel: ------------[ cut here ]------------ > > kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 > > kernel: Modules linked in: > > kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 > > kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) > > kernel: pc : __alloc_pages+0x1a0/0x290 > > kernel: lr : alloc_page_interleave+0x24/0xbc > > kernel: sp : ffff80000803bb90 > > kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 > > kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 > > kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e > > kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 > > kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 > > kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 > > kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 > > kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 > > kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 > > kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 > > kernel: Call trace: > > kernel: __alloc_pages+0x1a0/0x290 > > kernel: alloc_page_interleave+0x24/0xbc > > kernel: alloc_pages+0x10c/0x16c > > kernel: kmalloc_order+0x3c/0xc0 > > kernel: kmalloc_order_trace+0x38/0x130 > > kernel: __kmalloc+0x2e8/0x350 > > kernel: detect_cache_attributes+0x9c/0x110 > > kernel: update_siblings_masks+0x34/0x270 > > kernel: store_cpu_topology+0x7c/0xc0 > > kernel: smp_prepare_cpus+0x34/0x108 > > kernel: kernel_init_freeable+0x108/0x1b8 > > kernel: kernel_init+0x30/0x150 > > kernel: ret_from_fork+0x10/0x20 > > kernel: ---[ end trace 0000000000000000 ]--- > > kernel: Early cacheinfo failed, ret = -12 > > kernel: cblist_init_generic: Setting adjustable number of callback queues. > > kernel: cblist_init_generic: Setting shift to 2 and lim to 1. > > > > Thanks for the report. I will see if I can reproduce by forcing the absence > of PPTT which is something I haven't tested after I fixed hotplug issues > reported with my changes. OK I was able to reproduce this easily by just saying PPTT doesn't exist on my system. Not sure if my recent changes made this issue show up easily or it existed even w/o my changes but it is stupid issue :(. Can you check if this fixes it for you ? I can post the patch once you confirm. We assign negative/error value to unsigned int and then the number of leaves ends up with that value because of wrong comparison. Regards, Sudeep -->8 diff --git i/arch/arm64/kernel/cacheinfo.c w/arch/arm64/kernel/cacheinfo.c index 587543c6c51c..97c42be71338 100644 --- i/arch/arm64/kernel/cacheinfo.c +++ w/arch/arm64/kernel/cacheinfo.c @@ -45,7 +45,8 @@ static void ci_leaf_init(struct cacheinfo *this_leaf, int init_cache_level(unsigned int cpu) { - unsigned int ctype, level, leaves, fw_level; + unsigned int ctype, level, leaves; + int fw_level; struct cpu_cacheinfo *this_cpu_ci = get_cpu_cacheinfo(cpu); for (level = 1, leaves = 0; level <= MAX_CACHE_LEVEL; level++) { @@ -63,6 +64,9 @@ int init_cache_level(unsigned int cpu) else fw_level = acpi_find_last_cache_level(cpu); + if (fw_level < 0) + return fw_level; + if (level < fw_level) { /* * some external caches not specified in CLIDR_EL1 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 2022-08-05 15:08 ` Sudeep Holla @ 2022-08-05 18:38 ` Bruno Goncalves 0 siblings, 0 replies; 6+ messages in thread From: Bruno Goncalves @ 2022-08-05 18:38 UTC (permalink / raw) To: Sudeep Holla; +Cc: linux-arm-kernel, LKML, CKI Project On Fri, 5 Aug 2022 at 17:08, Sudeep Holla <sudeep.holla@arm.com> wrote: > Thanks for the fix, it works well. [ 0.003840] ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate [ 0.004588] cblist_init_generic: Setting adjustable number of callback queues. [ 0.004595] cblist_init_generic: Setting shift to 2 and lim to 1. [ 0.004707] cblist_init_generic: Setting shift to 2 and lim to 1. [ 0.004779] cblist_init_generic: Setting shift to 2 and lim to 1. [ 0.004908] rcu: Hierarchical SRCU implementation. Bruno > On Fri, Aug 05, 2022 at 02:59:55PM +0100, Sudeep Holla wrote: > > On Fri, Aug 05, 2022 at 11:47:57AM +0200, Bruno Goncalves wrote: > > > Hello, > > > > > > Since commit "c1c76700a0d6" we started to hit the following call trace > > > on aarch64 VMs when looking at journalctl log > > > > > > kernel: Mountpoint-cache hash table entries: 16384 (order: 5, 131072 > > > bytes, linear) > > > kernel: ACPI PPTT: No PPTT table found, CPU and cache topology may be inaccurate > > > kernel: ------------[ cut here ]------------ > > > kernel: WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 > > > kernel: Modules linked in: > > > kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0 #1 > > > kernel: pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) > > > kernel: pc : __alloc_pages+0x1a0/0x290 > > > kernel: lr : alloc_page_interleave+0x24/0xbc > > > kernel: sp : ffff80000803bb90 > > > kernel: x29: ffff80000803bb90 x28: 0000000000000000 x27: 0000000000000000 > > > kernel: x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 > > > kernel: x23: 000000000000001e x22: 0000000000000000 x21: 000000000000001e > > > kernel: x20: 000000000000001e x19: 0000000000040b20 x18: 0000000000000014 > > > kernel: x17: 6e69206562207961 x16: 6d2079676f6c6f70 x15: 6f74206568636163 > > > kernel: x14: 20646e6120555043 x13: 6574617275636361 x12: 6e69206562207961 > > > kernel: x11: 6d2079676f6c6f70 x10: 6f74206568636163 x9 : ffffa4f6617fd5e8 > > > kernel: x8 : ffffa4f663ed0440 x7 : 0000000000000007 x6 : 0000000000000000 > > > kernel: x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000 > > > kernel: x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffa4f663c53532 > > > kernel: Call trace: > > > kernel: __alloc_pages+0x1a0/0x290 > > > kernel: alloc_page_interleave+0x24/0xbc > > > kernel: alloc_pages+0x10c/0x16c > > > kernel: kmalloc_order+0x3c/0xc0 > > > kernel: kmalloc_order_trace+0x38/0x130 > > > kernel: __kmalloc+0x2e8/0x350 > > > kernel: detect_cache_attributes+0x9c/0x110 > > > kernel: update_siblings_masks+0x34/0x270 > > > kernel: store_cpu_topology+0x7c/0xc0 > > > kernel: smp_prepare_cpus+0x34/0x108 > > > kernel: kernel_init_freeable+0x108/0x1b8 > > > kernel: kernel_init+0x30/0x150 > > > kernel: ret_from_fork+0x10/0x20 > > > kernel: ---[ end trace 0000000000000000 ]--- > > > kernel: Early cacheinfo failed, ret = -12 > > > kernel: cblist_init_generic: Setting adjustable number of callback queues. > > > kernel: cblist_init_generic: Setting shift to 2 and lim to 1. > > > > > > > Thanks for the report. I will see if I can reproduce by forcing the absence > > of PPTT which is something I haven't tested after I fixed hotplug issues > > reported with my changes. > > OK I was able to reproduce this easily by just saying PPTT doesn't exist > on my system. Not sure if my recent changes made this issue show up easily > or it existed even w/o my changes but it is stupid issue :(. > > Can you check if this fixes it for you ? I can post the patch once > you confirm. We assign negative/error value to unsigned int and then > the number of leaves ends up with that value because of wrong comparison. > > Regards, > Sudeep > > -->8 > diff --git i/arch/arm64/kernel/cacheinfo.c w/arch/arm64/kernel/cacheinfo.c > index 587543c6c51c..97c42be71338 100644 > --- i/arch/arm64/kernel/cacheinfo.c > +++ w/arch/arm64/kernel/cacheinfo.c > @@ -45,7 +45,8 @@ static void ci_leaf_init(struct cacheinfo *this_leaf, > > int init_cache_level(unsigned int cpu) > { > - unsigned int ctype, level, leaves, fw_level; > + unsigned int ctype, level, leaves; > + int fw_level; > struct cpu_cacheinfo *this_cpu_ci = get_cpu_cacheinfo(cpu); > > for (level = 1, leaves = 0; level <= MAX_CACHE_LEVEL; level++) { > @@ -63,6 +64,9 @@ int init_cache_level(unsigned int cpu) > else > fw_level = acpi_find_last_cache_level(cpu); > > + if (fw_level < 0) > + return fw_level; > + > if (level < fw_level) { > /* > * some external caches not specified in CLIDR_EL1 > ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-08-05 18:38 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2022-08-05 9:47 [5.19.0] [aarch64] WARNING: CPU: 0 PID: 1 at mm/page_alloc.c:5407 __alloc_pages+0x1a0/0x290 Bruno Goncalves 2022-08-05 12:56 ` Mark Rutland 2022-08-05 13:28 ` Bruno Goncalves 2022-08-05 13:59 ` Sudeep Holla 2022-08-05 15:08 ` Sudeep Holla 2022-08-05 18:38 ` Bruno Goncalves
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).