On 2025-02-16 23:21:08 [-0500], Waiman Long wrote:
I would skip the first part. The backtrace is not really helpful here.
> The following bug report appeared with a test run in a RT debug kernel.
>
> [ 3359.353842] BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48
> [ 3359.353848] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 140605, name: kunit_try_catch
> [ 3359.353853] preempt_count: 1, expected: 0
> :
> [ 3359.353933] Call trace:
> :
> [ 3359.353955] rt_spin_lock+0x70/0x140
> [ 3359.353959] find_vmap_area+0x84/0x168
> [ 3359.353963] find_vm_area+0x1c/0x50
> [ 3359.353966] print_address_description.constprop.0+0x2a0/0x320
> [ 3359.353972] print_report+0x108/0x1f8
> [ 3359.353976] kasan_report+0x90/0xc8
> [ 3359.353980] __asan_load1+0x60/0x70
>
> Commit e30a0361b851 ("kasan: make report_lock a raw spinlock")
> changes report_lock to a raw_spinlock_t to avoid a similar RT problem.
s/to avoid.*//. This has nothing to do with the problem at hand.
> The print_address_description() function is called with report_lock
> acquired and interrupt disabled. However, the find_vm_area() function
> still needs to acquire a spinlock_t which becomes a sleeping lock in
> the RT kernel. IOW, we can't call find_vm_area() in a RT kernel and
> changing report_lock to a raw_spinlock_t is not enough to completely
> solve this RT kernel problem.
This function is always invoked under the report_lock which is a
raw_spinlock_t. The context under this lock is always atomic even on
PREEMPT_RT. find_vm_area() acquires vmap_node::busy.lock which is a
spinlock_t, becoming a sleeping lock on PREEMPT_RT and must not be
acquired in atomic context.
> Fix this bug report by skipping the find_vm_area() call in this case
> and just print out the address as is.
Please use PREEMPT_RT instead of RT.
Don't invoke find_vm_area() on PREEMPT_RT and just print the address.
Non-PREEMPT_RT builds remain unchanged. Add a DEFINE_WAIT_OVERRIDE_MAP()
is to tell lockdep that this lock nesting allowed because the PREEMPT_RT
part (which is invalid) has been taken care of.
> For !RT kernel, follow the example set in commit 0cce06ba859a
> ("debugobjects,locking: Annotate debug_object_fill_pool() wait type
> violation") and use DEFINE_WAIT_OVERRIDE_MAP() to avoid a spinlock_t
> inside raw_spinlock_t warning.
> Fixes: e30a0361b851 ("kasan: make report_lock a raw spinlock")
> Signed-off-by: Waiman Long <longman@xxxxxxxxxx>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> ---
> mm/kasan/report.c | 43 ++++++++++++++++++++++++++++++-------------
> 1 file changed, 30 insertions(+), 13 deletions(-)
>
> [v3] Rename helper to print_vmalloc_info_set_page.
>
> diff --git a/mm/kasan/report.c b/mm/kasan/report.c
> index 3fe77a360f1c..7c8c2e173aa4 100644
> --- a/mm/kasan/report.c
> +++ b/mm/kasan/report.c
> @@ -370,6 +370,34 @@ static inline bool init_task_stack_addr(const void *addr)
> sizeof(init_thread_union.stack));
> }
>
> +/*
> + * RT kernel cannot call find_vm_area() in atomic context. For !RT kernel,
> + * prevent spinlock_t inside raw_spinlock_t warning by raising wait-type
> + * to WAIT_SLEEP.
> + */
Do we need this comment? I lacks context of why it is atomic. And we
have it in the commit description.
> +static inline void print_vmalloc_info_set_page(void *addr, struct page **ppage)
> +{
> + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) {
> + static DEFINE_WAIT_OVERRIDE_MAP(vmalloc_map, LD_WAIT_SLEEP);
> + struct vm_struct *va;
> +
> + lock_map_acquire_try(&vmalloc_map);
> + va = find_vm_area(addr);
> + if (va) {
> + pr_err("The buggy address belongs to the virtual mapping at\n"
> + " [%px, %px) created by:\n"
> + " %pS\n",
> + va->addr, va->addr + va->size, va->caller);
> + pr_err("\n");
> +
> + *ppage = vmalloc_to_page(addr);
> + }
> + lock_map_release(&vmalloc_map);
> + return;
> + }
> + pr_err("The buggy address %px belongs to a vmalloc virtual mapping\n", addr);
> +}
> +
> static void print_address_description(void *addr, u8 tag,
> struct kasan_report_info *info)
> {
Sebastian
