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

[Tao Cui <cui.tao@linux.dev>]

于 2026年7月15日 GMT+08:00 23:01:09，David Laight <david.laight.linux@gmail.com> 写道：
>On Wed, 15 Jul 2026 20:44:19 +0800
>Tao Cui <cui.tao@linux.dev> wrote:
>
>> 在 2026/7/14 19:37, David Laight 写道:
>...
>> >>  	tg_update_slice(tg, rw);
>> >> @@ -1386,7 +1391,8 @@ static ssize_t tg_set_conf(struct kernfs_open_file *of,
>> >>  	if (is_u64)
>> >>  		*(u64 *)((void *)tg + of_cft(of)->private) = v;
>> >>  	else
>> >> -		*(unsigned int *)((void *)tg + of_cft(of)->private) = v;
>> >> +		*(unsigned int *)((void *)tg + of_cft(of)->private) =
>> >> +			min_t(u64, v, UINT_MAX);  
>> > 
>> > The LHS casts look horrid - there has to be a nicer way to do that.
>> > 
>> > And you don't need min_t() a plain min() will be fine.
>> >   
>> Hi David,
>> 
>> Both done in v2 — introduced a void *field local so the writes read
>> *(u64 *)field / *(unsigned int *)field, and switched to
>> min(v, (u64)UINT_MAX).
>
>You don't need the cast either.

Noted, this will be done in the next version.

Thanks,
Tao
>
>	David
>
>> 
>> Thanks,
>> Tao
>> > 	David
>> > 
>> > 
>> >   
>> >>  
>> >>  	tg_conf_updated(tg, false);
>> >>  	ret = 0;  
>> >   
>> 
>