Re: [PATCH] Fix a race in pid generation that causes pids to be reused immediately.

[Salman Qazi]

On Thu, Jun 10, 2010 at 1:38 PM,  <tytso@xxxxxxx> wrote:
> On Thu, Jun 10, 2010 at 01:09:11PM -0700, Salman wrote:
>> A program that repeatedly forks and waits is susceptible to having the
>> same pid repeated, especially when it competes with another instance of the
>> same program.  This is really bad for bash implementation.  Furthermore, many shell
>> scripts assume that pid numbers will not be used for some length of time.
>
> This should probably get wrapped at column 74 or so....
>
>> +static int pid_before(int base, int a, int b)
>> +{
>> +     /*
>> +      * This is the same as saying
>> +      *
>> +      * (a - base + MAXUINT) % MAXUINT < (b - base + MAXUINT) % MAXUINT
>> +      * and that mapping orders 'a' and 'b' with respect to 'base'.
>> +      *
>> +      */
>> +     return (unsigned)(a - base) < (unsigned)(b - base);
>> +}
>
> Does this work though if /proc/sys/kernel/pid_max is not set to
> MAXUINT?

Yes it does.  It should work for all values of pid_max.
>
> I like the optimization, but it looks like pid_max defaults to 4096 if
> CONFIG_BASE_SMALL is set, and 32768 otherwise.
>
> Am I missing something?

Yes.

(a - base + pid_max) % pid_max < (b - base + max_pid) % pid_max iff
(a - base + MAXUINT) % MAXUINT < (b - base + MAXUINT) % MAXUINT
for all pid_max <= MAXUINT.

The values of 'a' (or 'b')  in the range [base, pid_max) gets mapped
to [0, pid_max - base) and the range [0, base) gets mapped to
[MAXUINT, MAXUINT - base).  So, the order is essentially maintained by
this mapping.

>
>                                        - Ted
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/