Re: [PATCH v5] blk-throttle: fix divide-by-zero on legacy iops limit of 0

[yu kuai]

Hi,

在 2026/9/4 11:06, Tao Cui 写道:
> From: Tao Cui <cuitao@xxxxxxxxxx>
>
> Writing a multiple of 2^32 (e.g. 4294967296) to a legacy cgroup v1
> throttle iops file (blkio.throttle.{read,write}_iops_device) silently
> truncates to 0: tg_set_conf() stores the sscanf-parsed u64 value into
> an unsigned int field with no clamping. The cgroup v2 path,
> tg_set_limit(), already clamps the same kind of value with
> min_t(u64, val, UINT_MAX), but the legacy path never did. Note that
> the "!v -> U64_MAX" mapping only catches an explicit zero and does not
> catch a value that truncates to zero.
>
> With iops stored as 0, tg_update_has_rules() sets has_rules_iops[] and
> the next IO reaches tg_within_iops_limit(), which computes
>
>      jiffy_wait = max(jiffy_wait, HZ / iops_limit + 1);
>
> triggering a divide-by-zero oops.
>
> Fix it in tg_set_conf() by clamping the value to UINT_MAX, consistent
> with tg_set_limit(). This closes the truncation root cause: with 0 no
> longer reachable as a stored limit, the HZ / iops_limit divide is never
> hit.
>
> Signed-off-by: Tao Cui <cuitao@xxxxxxxxxx>
> Reviewed-by: Yu Kuai <yukuai@xxxxxxxxxx>
This tag is wrong, please use yukuai@xxxxxxx.
>
> ---
> Changes in v5:
> - Rebase onto current linux-next head (no code change, context shifted only).
> - Add the Reviewed-by tag collected on v4.
> - Link to v4: https://lore.kernel.org/r/20260722102459.253189-1-cui.tao@xxxxxxxxx
>
> Changes in v4:
> - Drop the defensive "iops_limit == 0" check in tg_dispatch_iops_time():
>    with the tg_set_conf() clamp in place, 0 can never be stored as a limit,
>    so the runtime check only guards an unreachable state. (Yu Kuai)
> - Drop the Fixes: tag: the unclamped write -- and the iops=0 behavior it
>    can produce (calculate_io_allowed() returns 0, so no IO is issued) --
>    long predates the commit that added the HZ / iops_limit divide, so
>    attributing it there was incorrect. (Yu Kuai)
>
> Changes in v3:
> - Drop the (u64) cast on UINT_MAX: the kernel's type-checked min() accepts
>    two unsigned types of different width (both >= 4 bytes), so
>    min(v, UINT_MAX) compiles clean. (David Laight)
>
> Changes in v2:
> - Use a "void *field" local for the config write so the assignment reads
>    *(u64 *)field / *(unsigned int *)field instead of the
>    (type *)((void *)tg + of_cft(of)->private) casts.
> - Use min(v, UINT_MAX) instead of min_t(u64, v, UINT_MAX).
> ---
>   block/blk-throttle.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/block/blk-throttle.c b/block/blk-throttle.c
> index ffc3b70065d4..97ad7959d006 100644
> --- a/block/blk-throttle.c
> +++ b/block/blk-throttle.c
> @@ -1383,10 +1383,12 @@ static ssize_t tg_set_conf(struct kernfs_open_file *of,
>   	tg = blkg_to_tg(ctx.blkg);
>   	tg_update_carryover(tg);
>   
> +	void *field = (void *)tg + of_cft(of)->private;
> +
>   	if (is_u64)
> -		*(u64 *)((void *)tg + of_cft(of)->private) = v;
> +		*(u64 *)field = v;
>   	else
> -		*(unsigned int *)((void *)tg + of_cft(of)->private) = v;
> +		*(unsigned int *)field = min(v, UINT_MAX);
>   
>   	tg_conf_updated(tg, false);
>   	ret = 0;

-- 
Thanks,
Kuai