Re: [RFC PATCH 2/2] mm: readahead: handle LARGE input to get_init_ra_size()

[Randy Dunlap <rdunlap@infradead.org>]

On 12/29/20 10:11 AM, Randy Dunlap wrote:
> On 12/29/20 10:01 AM, Toralf Förster wrote:
>> On 12/23/20 2:50 AM, Randy Dunlap wrote:
>>>> What motivates this change?  Is there any reason to think this can
>>>> happen?
>>> Spotted in the wild:
>> I run 2 hardened Gentoo systems, a server and a desktop.
>>
>> I patched the server with this:
>>
>> mr-fox ~ # cat ubsan.patch
>> --- linux-5.10.1.orig/mm/readahead.c
>> +++ linux-5.10.1/mm/readahead.c
>> @@ -310,7 +310,11 @@ void force_page_cache_ra(struct readahea
>>   */
>>  static unsigned long get_init_ra_size(unsigned long size, unsigned
>> long max)
>>  {
>> -       unsigned long newsize = roundup_pow_of_two(size);
>> +       unsigned long newsize;
>> +
>> +       if (!size)
>> +               size = 32;
>> +       newsize = roundup_pow_of_two(size);
>>
>>         if (newsize <= max / 32)
>>                 newsize = newsize * 4;
>>
>>
>>
>> and the issue did no longer occurred at the server (5.10.2).
>>
>> I did not patched the desktop system and the issue occurred still 3
>> times since 21th of december (5.10.2/3)
> 
> Yes, that's the patch that I posted on 2020-DEC-22.
> 
> Looks like I should submit a real patch for that.
> 
> thanks.
> 

Hi Toralf,

Do you want either or both of your
Reported-by: and Tested-by: on the patch?

thanks.
-- 
~Randy