| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <03b4e787-53bb-4a9c-afa5-dbb51599a8ce@arm.com>
Date: Mon, 13 Nov 2023 15:45:16 +0000
From: Robin Murphy <robin.murphy@....com>
To: kernel test robot <lkp@...el.com>, loongarch@...ts.linux.dev,
chenhuacai@...nel.org, kernel@...0n.name
Cc: oe-kbuild-all@...ts.linux.dev, linux-kernel@...r.kernel.org,
Will Deacon <will@...nel.org>,
Ilkka Koskinen <ilkka@...amperecomputing.com>
Subject: Re: drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse:
incorrect type in argument 1 (different address spaces)
On 2023-11-12 1:38 am, kernel test robot wrote:
> Hi Robin,
>
> First bad commit (maybe != root cause):
>
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head: 3ca112b71f35dd5d99fc4571a56b5fc6f0c15814
> commit: f9bd34e3753ea8f1433a3ba70f03a165a1416f98 perf/arm_cspmu: Clean up ACPI dependency
> date: 5 months ago
> config: loongarch-randconfig-r133-20231107 (https://download.01.org/0day-ci/archive/20231112/202311120926.cjYHyoYw-lkp@intel.com/config)
> compiler: loongarch64-linux-gcc (GCC) 13.2.0
> reproduce: (https://download.01.org/0day-ci/archive/20231112/202311120926.cjYHyoYw-lkp@intel.com/reproduce)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@...el.com>
> | Closes: https://lore.kernel.org/oe-kbuild-all/202311120926.cjYHyoYw-lkp@intel.com/
>
> sparse warnings: (new ones prefixed by >>)
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got unsigned int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got unsigned int [noderef] __percpu *
Unpicking the macros, I guess this must stem from:
for_each_sibling_event()
-> lockdep_assert_event_ctx()
->this_cpu_read(hardirqs_enabled)
which appears to be the only obvious place to involve a __percpu pointer
in this area.
From there it would seem to be a preexisting Loongarch bug - AFAICS
__percpu_read() (and presumably at least __percpu_write() as well)
should not be expecting a raw void * argument, but still the __percpu
pointer, since the final dereference is done right down in the asm using
the per-cpu offset stashed in r21?
Thanks,
Robin.
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got unsigned int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got unsigned int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got unsigned int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got unsigned int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got unsigned int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got unsigned int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got int [noderef] __percpu *
>>> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void *ptr @@ got int [noderef] __percpu * @@
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: expected void *ptr
> drivers/perf/arm_cspmu/arm_cspmu.c:627:9: sparse: got int [noderef] __percpu *
>
> vim +627 drivers/perf/arm_cspmu/arm_cspmu.c
>
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 609
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 610 /*
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 611 * Make sure the group of events can be scheduled at once
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 612 * on the PMU.
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 613 */
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 614 static bool arm_cspmu_validate_group(struct perf_event *event)
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 615 {
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 616 struct perf_event *sibling, *leader = event->group_leader;
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 617 struct arm_cspmu_hw_events fake_hw_events;
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 618
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 619 if (event->group_leader == event)
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 620 return true;
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 621
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 622 memset(&fake_hw_events, 0, sizeof(fake_hw_events));
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 623
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 624 if (!arm_cspmu_validate_event(event->pmu, &fake_hw_events, leader))
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 625 return false;
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 626
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 @627 for_each_sibling_event(sibling, leader) {
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 628 if (!arm_cspmu_validate_event(event->pmu, &fake_hw_events,
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 629 sibling))
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 630 return false;
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 631 }
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 632
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 633 return arm_cspmu_validate_event(event->pmu, &fake_hw_events, event);
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 634 }
> e37dfd65731dc4f Besar Wicaksono 2022-11-11 635
>
> :::::: The code at line 627 was first introduced by commit
> :::::: e37dfd65731dc4f001fa7dfa7f705e6840017d5a perf: arm_cspmu: Add support for ARM CoreSight PMU driver
>
> :::::: TO: Besar Wicaksono <bwicaksono@...dia.com>
> :::::: CC: Will Deacon <will@...nel.org>
>
Powered by blists - more mailing lists