ring-buffer: Sync IRQ works before buffer destruction

BugLink: https://bugs.launchpad.net/bugs/2025095

commit 675751b upstream.

If something was written to the buffer just before destruction,
it may be possible (maybe not in a real system, but it did
happen in ARCH=um with time-travel) to destroy the ringbuffer
before the IRQ work ran, leading this KASAN report (or a crash
without KASAN):

    BUG: KASAN: slab-use-after-free in irq_work_run_list+0x11a/0x13a
    Read of size 8 at addr 000000006d640a48 by task swapper/0

    CPU: 0 PID: 0 Comm: swapper Tainted: G        W  O       6.3.0-rc1 #7
    Stack:
     60c4f20f 0c203d48 41b58ab3 60f224fc
     600477fa 60f35687 60c4f20f 601273dd
     00000008 6101eb00 6101eab0 615be548
    Call Trace:
     [<60047a58>] show_stack+0x25e/0x282
     [<60c609e0>] dump_stack_lvl+0x96/0xfd
     [<60c50d4c>] print_report+0x1a7/0x5a8
     [<603078d3>] kasan_report+0xc1/0xe9
     [<60308950>] __asan_report_load8_noabort+0x1b/0x1d
     [<60232844>] irq_work_run_list+0x11a/0x13a
     [<602328b4>] irq_work_tick+0x24/0x34
     [<6017f9dc>] update_process_times+0x162/0x196
     [<6019f335>] tick_sched_handle+0x1a4/0x1c3
     [<6019fd9e>] tick_sched_timer+0x79/0x10c
     [<601812b9>] __hrtimer_run_queues.constprop.0+0x425/0x695
     [<60182913>] hrtimer_interrupt+0x16c/0x2c4
     [<600486a3>] um_timer+0x164/0x183
     [...]

    Allocated by task 411:
     save_stack_trace+0x99/0xb5
     stack_trace_save+0x81/0x9b
     kasan_save_stack+0x2d/0x54
     kasan_set_track+0x34/0x3e
     kasan_save_alloc_info+0x25/0x28
     ____kasan_kmalloc+0x8b/0x97
     __kasan_kmalloc+0x10/0x12
     __kmalloc+0xb2/0xe8
     load_elf_phdrs+0xee/0x182
     [...]

    The buggy address belongs to the object at 000000006d640800
     which belongs to the cache kmalloc-1k of size 1024
    The buggy address is located 584 bytes inside of
     freed 1024-byte region [000000006d640800, 000000006d640c00)

Add the appropriate irq_work_sync() so the work finishes before
the buffers are destroyed.

Prior to the commit in the Fixes tag below, there was only a
single global IRQ work, so this issue didn't exist.

Link: https://lore.kernel.org/linux-trace-kernel/20230427175920.a76159263122.I8295e405c44362a86c995e9c2c37e3e03810aa56@changeid

Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Fixes: 1569345 ("tracing/ring-buffer: Move poll wake ups into ring buffer code")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Roxana Nicolescu <roxana.nicolescu@canonical.com>

Author: jmberg-intel

kernel/trace/ring_buffer.c

@@ -1742,6 +1742,8 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
 	struct list_head *head = cpu_buffer->pages;
 	struct buffer_page *bpage, *tmp;
 
+	irq_work_sync(&cpu_buffer->irq_work.work);
+
 	free_buffer_page(cpu_buffer->reader_page);
 
 	if (head) {
@@ -1848,6 +1850,8 @@ ring_buffer_free(struct trace_buffer *buffer)
 
 	cpuhp_state_remove_instance(CPUHP_TRACE_RB_PREPARE, &buffer->node);
 
+	irq_work_sync(&buffer->irq_work.work);
+
 	for_each_buffer_cpu(buffer, cpu)
 		rb_free_cpu_buffer(buffer->buffers[cpu]);