| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <1bf686e4-6c22-47d9-d6e3-fd33faf59037@suse.cz>
Date: Thu, 27 Jan 2022 18:33:38 +0100
From: Vlastimil Babka <vbabka@...e.cz>
To: Liam Howlett <liam.howlett@...cle.com>
Cc: "maple-tree@...ts.infradead.org" <maple-tree@...ts.infradead.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Song Liu <songliubraving@...com>,
Davidlohr Bueso <dave@...olabs.net>,
"Paul E . McKenney" <paulmck@...nel.org>,
Matthew Wilcox <willy@...radead.org>,
Laurent Dufour <ldufour@...ux.ibm.com>,
David Rientjes <rientjes@...gle.com>,
Axel Rasmussen <axelrasmussen@...gle.com>,
Suren Baghdasaryan <surenb@...gle.com>,
Rik van Riel <riel@...riel.com>,
Peter Zijlstra <peterz@...radead.org>,
Michel Lespinasse <walken.cr@...il.com>,
Jerome Glisse <jglisse@...hat.com>,
Minchan Kim <minchan@...gle.com>,
Joel Fernandes <joelaf@...gle.com>,
Rom Lemarchand <romlem@...gle.com>
Subject: Re: [PATCH v4 55/66] mm/mempolicy: Use maple tree iterators instead
of vma linked list
On 1/27/22 18:25, Liam Howlett wrote:
> * Vlastimil Babka <vbabka@...e.cz> [220126 04:23]:
>> On 1/26/22 03:48, Liam Howlett wrote:
>> > * Vlastimil Babka <vbabka@...e.cz> [220120 06:58]:
>> >> On 12/1/21 15:30, Liam Howlett wrote:
>> >> > From: "Liam R. Howlett" <Liam.Howlett@...cle.com>
>> >> >
>> >> > Signed-off-by: Matthew Wilcox (Oracle) <willy@...radead.org>
>> >> > Signed-off-by: Liam R. Howlett <Liam.Howlett@...cle.com>
>> >> > ---
>> >> > mm/mempolicy.c | 53 ++++++++++++++++++++++++++------------------------
>> >> > 1 file changed, 28 insertions(+), 25 deletions(-)
>> >> >
>> >> > diff --git a/mm/mempolicy.c b/mm/mempolicy.c
>> >> > index 10e9c87260ed..0e2d918f4f30 100644
>> >> > --- a/mm/mempolicy.c
>> >> > +++ b/mm/mempolicy.c
>> >> > @@ -377,9 +377,10 @@ void mpol_rebind_task(struct task_struct *tsk, const nodemask_t *new)
>> >> > void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new)
>> >> > {
>> >> > struct vm_area_struct *vma;
>> >> > + MA_STATE(mas, &mm->mm_mt, 0, 0);
>> >>
>> >> VMA_ITERATOR?
>> >
>> > Yes, I will make this change. Thanks.
>> >
>> >>
>> >> >
>> >> > mmap_write_lock(mm);
>> >> > - for (vma = mm->mmap; vma; vma = vma->vm_next)
>> >> > + mas_for_each(&mas, vma, ULONG_MAX)
>> >> > mpol_rebind_policy(vma->vm_policy, new);
>> >> > mmap_write_unlock(mm);
>> >> > }
>> >> > @@ -652,7 +653,7 @@ static unsigned long change_prot_numa(struct vm_area_struct *vma,
>> >> > static int queue_pages_test_walk(unsigned long start, unsigned long end,
>> >> > struct mm_walk *walk)
>> >> > {
>> >> > - struct vm_area_struct *vma = walk->vma;
>> >> > + struct vm_area_struct *next, *vma = walk->vma;
>> >> > struct queue_pages *qp = walk->private;
>> >> > unsigned long endvma = vma->vm_end;
>> >> > unsigned long flags = qp->flags;
>> >> > @@ -667,9 +668,10 @@ static int queue_pages_test_walk(unsigned long start, unsigned long end,
>> >> > /* hole at head side of range */
>> >> > return -EFAULT;
>> >> > }
>> >> > + next = find_vma(vma->vm_mm, vma->vm_end);
>> >> > if (!(flags & MPOL_MF_DISCONTIG_OK) &&
>> >> > ((vma->vm_end < qp->end) &&
>> >> > - (!vma->vm_next || vma->vm_end < vma->vm_next->vm_start)))
>> >> > + (!next || vma->vm_end < next->vm_start)))
>> >> > /* hole at middle or tail of range */
>> >> > return -EFAULT;
>> >> >
>> >> > @@ -783,28 +785,24 @@ static int vma_replace_policy(struct vm_area_struct *vma,
>> >> > static int mbind_range(struct mm_struct *mm, unsigned long start,
>> >> > unsigned long end, struct mempolicy *new_pol)
>> >> > {
>> >> > - struct vm_area_struct *next;
>> >> > + MA_STATE(mas, &mm->mm_mt, start - 1, start - 1);
>> >> > struct vm_area_struct *prev;
>> >> > struct vm_area_struct *vma;
>> >> > int err = 0;
>> >> > pgoff_t pgoff;
>> >> > - unsigned long vmstart;
>> >> > - unsigned long vmend;
>> >> > -
>> >> > - vma = find_vma(mm, start);
>> >> > - VM_BUG_ON(!vma);
>> >> >
>> >> > - prev = vma->vm_prev;
>> >> > - if (start > vma->vm_start)
>> >> > - prev = vma;
>> >> > + prev = mas_find_rev(&mas, 0);
>> >> > + if (prev && (start < prev->vm_end))
>> >> > + vma = prev;
>> >> > + else
>> >> > + vma = mas_next(&mas, end - 1);
>> >> >
>> >> > - for (; vma && vma->vm_start < end; prev = vma, vma = next) {
>> >> > - next = vma->vm_next;
>> >> > - vmstart = max(start, vma->vm_start);
>> >> > - vmend = min(end, vma->vm_end);
>> >> > + do {
>> >> > + unsigned long vmstart = max(start, vma->vm_start);
>> >> > + unsigned long vmend = min(end, vma->vm_end);
>> >>
>> >> What if vma is null? Shouldn't this rather be a "for (; vma; vma =
>> >> mas_next(...)"
>> >
>> > We have already found the vma above.
>>
>> AFAICS if the range intersects no vmas, we might have found a 'prev', but
>> 'vma' might be NULL after the "vma = mas_next(&mas, end - 1);"?
>
> Yes, I agree. I was going to restore VM_BUG_ON(!vma) after mas_next(),
> but that's not the same as the existing code. The VM_BUG_ON(!vma)
> only catches if we search for 'start' above any VMA. It seems
> mbind_range() would have returned without error if there were no VMAs
> within the range but would error if the 'start' was sufficiently large.
Ah I missed there was a VM_BUG_ON(!vma) previously and that the callers seem
to only call mbind_range() if there's an actual vma in the range, so I guess
my suggestion was misguided.
> I think it is better to write it as you suggested to restore the
> functionality of not failing on an empty list. I don't see a decent
> way of checking if we searched for an address above any VMA to restore
> the VM_BUG_ON() which existed - although I see little value in it to
> begin with. I will mention this in the commit message.
Powered by blists - more mailing lists