[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJZ5v0h82YpaG9a1mmavVnm4WwjDV5iU_Qof6dN_PqP1UR-OUQ@mail.gmail.com>
Date: Thu, 3 Nov 2022 17:57:04 +0100
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: TGSP <tgsp002@...il.com>
Cc: rafael@...nel.org, len.brown@...el.com, pavel@....cz,
huanglei@...inos.cn, linux-pm@...r.kernel.org,
linux-kernel@...r.kernel.org, xiongxin <xiongxin@...inos.cn>,
stable@...r.kernel.org
Subject: Re: [PATCH -next 2/2] PM: hibernate: add check of preallocate mem for
image size pages
On Thu, Nov 3, 2022 at 5:47 PM Rafael J. Wysocki <rafael@...nel.org> wrote:
>
> On Tue, Nov 1, 2022 at 3:28 AM TGSP <tgsp002@...il.com> wrote:
> >
> > From: xiongxin <xiongxin@...inos.cn>
> >
> > Added a check on the return value of preallocate_image_highmem(). If
> > memory preallocate is insufficient, S4 cannot be done;
> >
> > I am playing 4K video on a machine with AMD or other graphics card and
> > only 8GiB memory, and the kernel is not configured with CONFIG_HIGHMEM.
> > When doing the S4 test, the analysis found that when the pages get from
> > minimum_image_size() is large enough, The preallocate_image_memory() and
> > preallocate_image_highmem() calls failed to obtain enough memory. Add
> > the judgment that memory preallocate is insufficient;
> >
> > "pages -= free_unnecessary_pages()" below will let pages to drop a lot,
> > so I wonder if it makes sense to add a judgment here.
> >
> > Cc: stable@...r.kernel.org
> > Signed-off-by: xiongxin <xiongxin@...inos.cn>
> > Signed-off-by: huanglei <huanglei@...inos.cn>
> > ---
> > kernel/power/snapshot.c | 9 ++++++++-
> > 1 file changed, 8 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c
> > index c20ca5fb9adc..670abf89cf31 100644
> > --- a/kernel/power/snapshot.c
> > +++ b/kernel/power/snapshot.c
> > @@ -1738,6 +1738,7 @@ int hibernate_preallocate_memory(void)
> > struct zone *zone;
> > unsigned long saveable, size, max_size, count, highmem, pages = 0;
> > unsigned long alloc, save_highmem, pages_highmem, avail_normal;
> > + unsigned long size_highmem;
>
> Please define this in the block where it will be used.
>
> > ktime_t start, stop;
> > int error;
> >
> > @@ -1863,7 +1864,13 @@ int hibernate_preallocate_memory(void)
> > pages_highmem += size;
>
> The line above can be dropped.
Not really, it needs to be replaced with
size_highmem = size ;
> > alloc -= size;
> > size = preallocate_image_memory(alloc, avail_normal);
> > - pages_highmem += preallocate_image_highmem(alloc - size);
> > + size_highmem += preallocate_image_highmem(alloc - size);
>
> Did you mean "="?
>
> Assuming you did, this could be
>
> size_highmem = size + preallocate_image_highmem(alloc - size);
And here
size_highmem += preallocate_image_highmem(alloc - size);
which is what you had in the original patch.
> if (size_highmem < alloc) {
>
> > + if (size_highmem < (alloc - size)) {
>
> The inner parens were not necessary.
>
> > + pr_err("Image allocation is %lu pages short, exit\n",
> > + alloc - size - pages_highmem);
> > + goto err_out;
> > + }
> > + pages_highmem += size_highmem;
> > pages += pages_highmem + size;
> > }
> >
> > --
>
> But overall it would be better to avoid bailing out.
Powered by blists - more mailing lists