Quoting Wei Yongjun (2020-08-17 13:28:26)
> kmemleak report memory leak as follows:
>
> unreferenced object 0x607ee4e5f948 (size 8):
> comm "syz-executor.1", pid 2098, jiffies 4295031601 (age 288.468s)
> hex dump (first 8 bytes):
> 00 00 00 00 00 00 00 00
> backtrace:
> [] relay_open kernel/relay.c:583 [inline]
> [] relay_open+0xb6/0x970 kernel/relay.c:563
> [<38ae5a4b>] do_blk_trace_setup+0x4a8/0xb20
> kernel/trace/blktrace.c:557
> [] __blk_trace_setup+0xb6/0x150 kernel/trace/blktrace.c:597
> [<38fdf803>] blk_trace_ioctl+0x146/0x280 kernel/trace/blktrace.c:738
> [] blkdev_ioctl+0xb2/0x6a0 block/ioctl.c:613
> [<579e47e0>] block_ioctl+0xe5/0x120 fs/block_dev.c:1871
> [] vfs_ioctl fs/ioctl.c:48 [inline]
> [] __do_sys_ioctl fs/ioctl.c:753 [inline]
> [] __se_sys_ioctl fs/ioctl.c:739 [inline]
> [] __x64_sys_ioctl+0x170/0x1ce fs/ioctl.c:739
> [<88fc9942>] do_syscall_64+0x33/0x40 arch/x86/entry/common.c:46
> [<4f6dd57a>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
>
> 'chan->buf' is malloced in relay_open() by alloc_percpu() but not free
> while destroy the relay channel. Fix it by adding free_percpu() before
> return from relay_destroy_channel().
>
> Fixes: 017c59c042d0 ("relay: Use per CPU constructs for the relay channel
> buffer pointers")
> Reported-by: Hulk Robot
> Signed-off-by: Wei Yongjun
> ---
> kernel/relay.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/kernel/relay.c b/kernel/relay.c
> index 72fe443ea78f..fb4e0c530c08 100644
> --- a/kernel/relay.c
> +++ b/kernel/relay.c
> @@ -197,6 +197,7 @@ static struct rchan_buf *relay_create_buf(struct rchan
> *chan)
> static void relay_destroy_channel(struct kref *kref)
> {
> struct rchan *chan = container_of(kref, struct rchan, kref);
> + free_percpu(chan->buf);
> kfree(chan);
That catches the error path for relay_open as well, and the fixes is
correct. Worth a cc:stable, #v4.9+?
Reviewed-by: Chris Wilson
-Chris
