On 2020/03/13 18:46, Tetsuo Handa wrote:
> On 2020/03/12 23:38, Al Viro wrote:
>> It _does_ look like that double-fput() is real, but
>> I'd like a confirmation before going further - umh is convoluted
>> enough for something subtle to be hidden there. Alexei, what
>> the refcounting behaviour was supposed to be? As in "this
>> function consumes the reference passed to it in this argument",
>> etc.
>>
>
> Yes, double-fput() is easily observable as POISON_FREE pattern
> using debug printk() patch and sample kernel module shown below.
>
No response from Alexei, but I think that 449325b52b7a6208 ("umh:
introduce fork_usermode_blob() helper") just did not realize that
opening a file for execution needs special handling (i.e. denying
write access) compared to opening a file for read or write.
Can we send this patch?
