Re: [Xen-devel] [PATCH v4 1/4] x86/mem_sharing: reorder when pages are unlocked and released

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 03.05.19 at 15:48, <tamas@xxxxxxxxxxxxx> wrote:
> On Fri, May 3, 2019 at 2:12 AM Jan Beulich <JBeulich@xxxxxxxx> wrote:
>>
>> >>> On 03.05.19 at 00:13, <tamas@xxxxxxxxxxxxx> wrote:
>> > @@ -1002,7 +989,10 @@ static int share_pages(struct domain *sd, gfn_t 
>> > sgfn, shr_handle_t sh,
>> >      /* Free the client page */
>> >      if(test_and_clear_bit(_PGC_allocated, &cpage->count_info))
>> >          put_page(cpage);

This should be after ...

>> > -    put_page(cpage);
>> > +
>> > +    BUG_ON(!put_count);

... this, because ...

>> > +    while ( put_count-- )
>> > +        put_page_and_type(cpage);
>>
>> Strictly speaking I think the BUG_ON() should be moved ahead of the
>> if() in context, so that a problematic put_page() would not get
>> executed in the first place (even if the system is to die soon after).
> 
> I don't follow - where is the problematic put_page()? And why is it 
> problematic?

... if indeed the BUG_ON() triggers, then it should do so before
potentially putting the last ref of a page which shouldn't be put
that way.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel