The patch titled Subject: s390/irqflags: ddo not instrument arch_local_irq_*() with KMSAN has been added to the -mm mm-unstable branch. Its filename is s390-irqflags-do-not-instrument-arch_local_irq_-with-kmsan.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/s390-irqflags-do-not-instrument-arch_local_irq_-with-kmsan.patch This patch will later appear in the mm-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Ilya Leoshkevich <iii@xxxxxxxxxxxxx> Subject: s390/irqflags: ddo not instrument arch_local_irq_*() with KMSAN Date: Thu, 14 Dec 2023 00:24:47 +0100 KMSAN generates the following false positives on s390x: [ 6.063666] DEBUG_LOCKS_WARN_ON(lockdep_hardirqs_enabled()) [ ...] [ 6.577050] Call Trace: [ 6.619637] [<000000000690d2de>] check_flags+0x1fe/0x210 [ 6.665411] ([<000000000690d2da>] check_flags+0x1fa/0x210) [ 6.707478] [<00000000006cec1a>] lock_acquire+0x2ca/0xce0 [ 6.749959] [<00000000069820ea>] _raw_spin_lock_irqsave+0xea/0x190 [ 6.794912] [<00000000041fc988>] __stack_depot_save+0x218/0x5b0 [ 6.838420] [<000000000197affe>] __msan_poison_alloca+0xfe/0x1a0 [ 6.882985] [<0000000007c5827c>] start_kernel+0x70c/0xd50 [ 6.927454] [<0000000000100036>] startup_continue+0x36/0x40 Between trace_hardirqs_on() and `stosm __mask, 3` lockdep thinks that interrupts are on, but on the CPU they are still off. KMSAN instrumentation takes spinlocks, giving lockdep a chance to see and complain about this discrepancy. KMSAN instrumentation is inserted in order to poison the __mask variable. Disable instrumentation in the respective functions. They are very small and it's easy to see that no important metadata updates are lost because of this. Link: https://lkml.kernel.org/r/20231213233605.661251-28-iii@xxxxxxxxxxxxx Signed-off-by: Ilya Leoshkevich <iii@xxxxxxxxxxxxx> Cc: Alexander Gordeev <agordeev@xxxxxxxxxxxxx> Cc: Alexander Potapenko <glider@xxxxxxxxxx> Cc: Christian Borntraeger <borntraeger@xxxxxxxxxxxxx> Cc: Christoph Lameter <cl@xxxxxxxxx> Cc: David Rientjes <rientjes@xxxxxxxxxx> Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx> Cc: Heiko Carstens <hca@xxxxxxxxxxxxx> Cc: Hyeonggon Yoo <42.hyeyoo@xxxxxxxxx> Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx> Cc: Marco Elver <elver@xxxxxxxxxx> Cc: Mark Rutland <mark.rutland@xxxxxxx> Cc: Masami Hiramatsu <mhiramat@xxxxxxxxxx> Cc: Pekka Enberg <penberg@xxxxxxxxxx> Cc: Roman Gushchin <roman.gushchin@xxxxxxxxx> Cc: Steven Rostedt (Google) <rostedt@xxxxxxxxxxx> Cc: Sven Schnelle <svens@xxxxxxxxxxxxx> Cc: Vasily Gorbik <gor@xxxxxxxxxxxxx> Cc: Vlastimil Babka <vbabka@xxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- arch/s390/include/asm/irqflags.h | 18 +++++++++++++++--- drivers/s390/char/sclp.c | 2 +- 2 files changed, 16 insertions(+), 4 deletions(-) --- a/arch/s390/include/asm/irqflags.h~s390-irqflags-do-not-instrument-arch_local_irq_-with-kmsan +++ a/arch/s390/include/asm/irqflags.h @@ -37,12 +37,19 @@ static __always_inline void __arch_local asm volatile("ssm %0" : : "Q" (flags) : "memory"); } -static __always_inline unsigned long arch_local_save_flags(void) +#ifdef CONFIG_KMSAN +#define ARCH_LOCAL_IRQ_ATTRIBUTES \ + noinline notrace __no_sanitize_memory __maybe_unused +#else +#define ARCH_LOCAL_IRQ_ATTRIBUTES __always_inline +#endif + +static ARCH_LOCAL_IRQ_ATTRIBUTES unsigned long arch_local_save_flags(void) { return __arch_local_irq_stnsm(0xff); } -static __always_inline unsigned long arch_local_irq_save(void) +static ARCH_LOCAL_IRQ_ATTRIBUTES unsigned long arch_local_irq_save(void) { return __arch_local_irq_stnsm(0xfc); } @@ -52,7 +59,12 @@ static __always_inline void arch_local_i arch_local_irq_save(); } -static __always_inline void arch_local_irq_enable(void) +static ARCH_LOCAL_IRQ_ATTRIBUTES void arch_local_irq_enable_external(void) +{ + __arch_local_irq_stosm(0x01); +} + +static ARCH_LOCAL_IRQ_ATTRIBUTES void arch_local_irq_enable(void) { __arch_local_irq_stosm(0x03); } --- a/drivers/s390/char/sclp.c~s390-irqflags-do-not-instrument-arch_local_irq_-with-kmsan +++ a/drivers/s390/char/sclp.c @@ -736,7 +736,7 @@ sclp_sync_wait(void) cr0_sync.val = cr0.val & ~CR0_IRQ_SUBCLASS_MASK; cr0_sync.val |= 1UL << (63 - 54); local_ctl_load(0, &cr0_sync); - __arch_local_irq_stosm(0x01); + arch_local_irq_enable_external(); /* Loop until driver state indicates finished request */ while (sclp_running_state != sclp_running_state_idle) { /* Check for expired request timer */ _ Patches currently in -mm which might be from iii@xxxxxxxxxxxxx are ftrace-unpoison-ftrace_regs-in-ftrace_ops_list_func.patch kmsan-make-the-tests-compatible-with-kmsanpanic=1.patch kmsan-disable-kmsan-when-deferred_struct_page_init-is-enabled.patch kmsan-increase-the-maximum-store-size-to-4096.patch kmsan-fix-is_bad_asm_addr-on-arches-with-overlapping-address-spaces.patch kmsan-fix-kmsan_copy_to_user-on-arches-with-overlapping-address-spaces.patch kmsan-remove-a-useless-assignment-from-kmsan_vmap_pages_range_noflush.patch kmsan-remove-an-x86-specific-include-from-kmsanh.patch kmsan-expose-kmsan_get_metadata.patch kmsan-export-panic_on_kmsan.patch kmsan-allow-disabling-kmsan-checks-for-the-current-task.patch kmsan-support-slab_poison.patch kmsan-use-align_down-in-kmsan_get_metadata.patch mm-slub-let-kmsan-access-metadata.patch mm-slub-unpoison-the-memchr_inv-return-value.patch mm-kfence-disable-kmsan-when-checking-the-canary.patch lib-zlib-unpoison-dfltcc-output-buffers.patch kmsan-accept-ranges-starting-with-0-on-s390.patch s390-turn-off-kmsan-for-boot-vdso-and-purgatory.patch s390-use-a-larger-stack-for-kmsan.patch s390-boot-add-the-kmsan-runtime-stub.patch s390-checksum-add-a-kmsan-check.patch s390-cpacf-unpoison-the-results-of-cpacf_trng.patch s390-cpumf-unpoison-stcctm-output-buffer.patch s390-diag-unpoison-diag224-output-buffer.patch s390-ftrace-unpoison-ftrace_regs-in-kprobe_ftrace_handler.patch s390-irqflags-do-not-instrument-arch_local_irq_-with-kmsan.patch s390-mm-define-kmsan-metadata-for-vmalloc-and-modules.patch s390-string-add-kmsan-support.patch s390-traps-unpoison-the-kernel_stack_overflows-pt_regs.patch s390-uaccess-add-kmsan-support-to-put_user-and-get_user.patch s390-unwind-disable-kmsan-checks.patch s390-implement-the-architecture-specific-kmsan-functions.patch kmsan-enable-on-s390.patch