From: Suren Baghdasaryan <surenb@google.com>
To: "Liam R. Howlett" <Liam.Howlett@oracle.com>,
Suren Baghdasaryan <surenb@google.com>,
akpm@linux-foundation.org, torvalds@linux-foundation.org,
jannh@google.com, willy@infradead.org, david@redhat.com,
peterx@redhat.com, ldufour@linux.ibm.com, vbabka@suse.cz,
michel@lespinasse.org, jglisse@google.com, mhocko@suse.com,
hannes@cmpxchg.org, dave@stgolabs.net, hughd@google.com,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
stable@vger.kernel.org,
Linus Torvalds <torvalds@linuxfoundation.org>
Subject: Re: [PATCH v2 6/6] mm: move vma locking out of vma_prepare
Date: Wed, 2 Aug 2023 10:24:31 -0700 [thread overview]
Message-ID: <CAJuCfpF-E65wJcprA80zLWfbfHdaDJMafNe087aUs_-SkwUTdA@mail.gmail.com> (raw)
In-Reply-To: <20230802165900.otlirhbjr54wymvl@revolver>
On Wed, Aug 2, 2023 at 9:59 AM Liam R. Howlett <Liam.Howlett@oracle.com> wrote:
>
> * Suren Baghdasaryan <surenb@google.com> [230801 18:07]:
> > vma_prepare() is currently the central place where vmas are being locked
> > before vma_complete() applies changes to them. While this is convenient,
> > it also obscures vma locking and makes it hard to follow the locking rules.
> > Move vma locking out of vma_prepare() and take vma locks explicitly at the
> > locations where vmas are being modified.
> >
> > Suggested-by: Linus Torvalds <torvalds@linuxfoundation.org>
> > Signed-off-by: Suren Baghdasaryan <surenb@google.com>
> > Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
> > ---
> > mm/mmap.c | 26 ++++++++++++++++----------
> > 1 file changed, 16 insertions(+), 10 deletions(-)
> >
> > diff --git a/mm/mmap.c b/mm/mmap.c
> > index 850a39dee075..e59d83cb1d7a 100644
> > --- a/mm/mmap.c
> > +++ b/mm/mmap.c
> > @@ -476,16 +476,6 @@ static inline void init_vma_prep(struct vma_prepare *vp,
> > */
> > static inline void vma_prepare(struct vma_prepare *vp)
> > {
> > - vma_start_write(vp->vma);
> > - if (vp->adj_next)
> > - vma_start_write(vp->adj_next);
> > - if (vp->insert)
> > - vma_start_write(vp->insert);
> > - if (vp->remove)
> > - vma_start_write(vp->remove);
> > - if (vp->remove2)
> > - vma_start_write(vp->remove2);
> > -
> > if (vp->file) {
> > uprobe_munmap(vp->vma, vp->vma->vm_start, vp->vma->vm_end);
> >
> > @@ -650,6 +640,7 @@ int vma_expand(struct vma_iterator *vmi, struct vm_area_struct *vma,
> > bool remove_next = false;
> > struct vma_prepare vp;
> >
> > + vma_start_write(vma);
>
> This lock made me think about the lock in dup_anon_vma().. the only
> reason we need to dup the anon vma is because the VMA _will_ be
> modified.. So if we are to remove next the dup_anon_vma() call may lock
> vma again. This happens multiple times through this patch set.
>
> If we lock both dst and src VMAs before calling dup_anon_vma, then I
> think it will become more clear in the code below which VMAs are locked.
> We could remove the lock in the dup_anon_vma() (and replace with an
> assert?), and drop the vma locking of "vma".
>
> I think this would address the confusion of the locking below that I
> experienced and avoid calling the vma_start_write() multiple times for
> the same vma. I mean, having the vma_start_write(vma) obscures which
> VMA is locked as well.
Agree, it would result in an easier to understand locking code. I'll
move the lock outside of dup_anon_vma() and will replace it with an
assert. Thanks!
>
>
> > if (next && (vma != next) && (end == next->vm_end)) {
> > int ret;
> >
> > @@ -657,6 +648,7 @@ int vma_expand(struct vma_iterator *vmi, struct vm_area_struct *vma,
> > ret = dup_anon_vma(vma, next);
> > if (ret)
> > return ret;
> > + vma_start_write(next);
> > }
> >
> > init_multi_vma_prep(&vp, vma, NULL, remove_next ? next : NULL, NULL);
> > @@ -708,6 +700,8 @@ int vma_shrink(struct vma_iterator *vmi, struct vm_area_struct *vma,
> > if (vma_iter_prealloc(vmi))
> > return -ENOMEM;
> >
> > + vma_start_write(vma);
> > +
> > init_vma_prep(&vp, vma);
> > vma_prepare(&vp);
> > vma_adjust_trans_huge(vma, start, end, 0);
> > @@ -946,10 +940,12 @@ struct vm_area_struct *vma_merge(struct vma_iterator *vmi, struct mm_struct *mm,
> > /* Can we merge both the predecessor and the successor? */
> > if (merge_prev && merge_next &&
> > is_mergeable_anon_vma(prev->anon_vma, next->anon_vma, NULL)) {
> > + vma_start_write(next);
> > remove = next; /* case 1 */
> > vma_end = next->vm_end;
> > err = dup_anon_vma(prev, next);
> > if (curr) { /* case 6 */
> > + vma_start_write(curr);
> > remove = curr;
> > remove2 = next;
> > if (!next->anon_vma)
> > @@ -958,6 +954,7 @@ struct vm_area_struct *vma_merge(struct vma_iterator *vmi, struct mm_struct *mm,
> > } else if (merge_prev) { /* case 2 */
> > if (curr) {
> > err = dup_anon_vma(prev, curr);
> > + vma_start_write(curr);
> > if (end == curr->vm_end) { /* case 7 */
> > remove = curr;
> > } else { /* case 5 */
> > @@ -969,6 +966,7 @@ struct vm_area_struct *vma_merge(struct vma_iterator *vmi, struct mm_struct *mm,
> > res = next;
> > if (prev && addr < prev->vm_end) { /* case 4 */
> > vma_end = addr;
> > + vma_start_write(next);
> > adjust = next;
> > adj_start = -(prev->vm_end - addr);
> > err = dup_anon_vma(next, prev);
> > @@ -983,6 +981,7 @@ struct vm_area_struct *vma_merge(struct vma_iterator *vmi, struct mm_struct *mm,
> > vma_pgoff = next->vm_pgoff - pglen;
> > if (curr) { /* case 8 */
> > vma_pgoff = curr->vm_pgoff;
> > + vma_start_write(curr);
> > remove = curr;
> > err = dup_anon_vma(next, curr);
> > }
> > @@ -996,6 +995,8 @@ struct vm_area_struct *vma_merge(struct vma_iterator *vmi, struct mm_struct *mm,
> > if (vma_iter_prealloc(vmi))
> > return NULL;
> >
> > + vma_start_write(vma);
> > +
> > init_multi_vma_prep(&vp, vma, adjust, remove, remove2);
> > VM_WARN_ON(vp.anon_vma && adjust && adjust->anon_vma &&
> > vp.anon_vma != adjust->anon_vma);
> > @@ -2373,6 +2374,9 @@ int __split_vma(struct vma_iterator *vmi, struct vm_area_struct *vma,
> > if (new->vm_ops && new->vm_ops->open)
> > new->vm_ops->open(new);
> >
> > + vma_start_write(vma);
> > + vma_start_write(new);
> > +
> > init_vma_prep(&vp, vma);
> > vp.insert = new;
> > vma_prepare(&vp);
> > @@ -3078,6 +3082,8 @@ static int do_brk_flags(struct vma_iterator *vmi, struct vm_area_struct *vma,
> > if (vma_iter_prealloc(vmi))
> > goto unacct_fail;
> >
> > + vma_start_write(vma);
> > +
> > init_vma_prep(&vp, vma);
> > vma_prepare(&vp);
> > vma_adjust_trans_huge(vma, vma->vm_start, addr + len, 0);
> > --
> > 2.41.0.585.gd2178a4bd4-goog
> >
next prev parent reply other threads:[~2023-08-02 17:24 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-01 22:07 [PATCH v2 0/6] make vma locking more obvious Suren Baghdasaryan
2023-08-01 22:07 ` [PATCH v2 1/6] mm: enable page walking API to lock vmas during the walk Suren Baghdasaryan
2023-08-01 22:07 ` [PATCH v2 2/6] mm: for !CONFIG_PER_VMA_LOCK equate write lock assertion for vma and mmap Suren Baghdasaryan
2023-08-02 17:07 ` Liam R. Howlett
2023-08-01 22:07 ` [PATCH v2 3/6] mm: replace mmap with vma write lock assertions when operating on a vma Suren Baghdasaryan
2023-08-02 17:13 ` Liam R. Howlett
2023-08-01 22:07 ` [PATCH v2 4/6] mm: lock vma explicitly before doing vm_flags_reset and vm_flags_reset_once Suren Baghdasaryan
2023-08-02 17:13 ` Liam R. Howlett
2023-08-02 17:49 ` Linus Torvalds
2023-08-02 18:09 ` Suren Baghdasaryan
2023-08-02 18:53 ` Linus Torvalds
2023-08-02 20:21 ` Suren Baghdasaryan
2023-08-01 22:07 ` [PATCH v2 5/6] mm: always lock new vma before inserting into vma tree Suren Baghdasaryan
2023-08-02 17:02 ` Liam R. Howlett
2023-08-01 22:07 ` [PATCH v2 6/6] mm: move vma locking out of vma_prepare Suren Baghdasaryan
2023-08-02 16:59 ` Liam R. Howlett
2023-08-02 17:24 ` Suren Baghdasaryan [this message]
2023-08-03 17:28 ` [PATCH v2 0/6] make vma locking more obvious Suren Baghdasaryan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAJuCfpF-E65wJcprA80zLWfbfHdaDJMafNe087aUs_-SkwUTdA@mail.gmail.com \
--to=surenb@google.com \
--cc=Liam.Howlett@oracle.com \
--cc=akpm@linux-foundation.org \
--cc=dave@stgolabs.net \
--cc=david@redhat.com \
--cc=hannes@cmpxchg.org \
--cc=hughd@google.com \
--cc=jannh@google.com \
--cc=jglisse@google.com \
--cc=ldufour@linux.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=michel@lespinasse.org \
--cc=peterx@redhat.com \
--cc=stable@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=torvalds@linuxfoundation.org \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).