From: Yafang Shao <laoar.shao@gmail.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Linux MM <linux-mm@kvack.org>,
"Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>
Subject: Re: [PATCH] mm, gpu: fix error when FOLL_MLOCK an unpresent page
Date: Mon, 30 Aug 2021 17:12:32 +0800 [thread overview]
Message-ID: <CALOAHbBocrx-hTe57Rb=ipOQaN78+syqeb1tbuj72Xak+9wUZw@mail.gmail.com> (raw)
In-Reply-To: <YSyCDxWWvKP2OthC@infradead.org>
On Mon, Aug 30, 2021 at 3:00 PM Christoph Hellwig <hch@infradead.org> wrote:
>
> On Sun, Aug 29, 2021 at 01:19:53AM +0000, Yafang Shao wrote:
> > After some analyzation, I found it was caused by a bug in GUP.
> > When the kernel module calls get_user_pages() with FOLL_MLOCK being set but
> > FOLL_POPULATE being unset, if the page of the user addr isn't present, the
>
> Which is not a valid way to call get_user_pages. What we need to do is
> to reject that case.
Do you mean below change ?
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 7ca22e6e694a..10f7d6f2ad7b 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2855,7 +2855,7 @@ struct page *follow_page(struct vm_area_struct
*vma, unsigned long address,
#define FOLL_NUMA 0x200 /* force NUMA hinting page fault */
#define FOLL_MIGRATION 0x400 /* wait for page to replace migration entry */
#define FOLL_TRIED 0x800 /* a retry, previous pass started an IO */
-#define FOLL_MLOCK 0x1000 /* lock present pages */
+#define FOLL_MLOCK 0x1000 /* lock present pages, must be set
with FOLL_POPULATE */
#define FOLL_REMOTE 0x2000 /* we are working on non-current tsk/mm */
#define FOLL_COW 0x4000 /* internal GUP flag */
#define FOLL_ANON 0x8000 /* don't do file mappings */
diff --git a/mm/gup.c b/mm/gup.c
index b94717977d17..dfdc0654f7a5 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -929,9 +929,9 @@ static int faultin_page(struct vm_area_struct *vma,
unsigned int fault_flags = 0;
vm_fault_t ret;
- /* mlock all present pages, but do not fault in new pages */
- if ((*flags & (FOLL_POPULATE | FOLL_MLOCK)) == FOLL_MLOCK)
- return -ENOENT;
+ /* FOLL_MLOCK must be set with FOLL_POPULATE */
+ BUG_ON((*flags & (FOLL_POPULATE | FOLL_MLOCK)) == FOLL_MLOCK);
+
if (*flags & FOLL_WRITE)
fault_flags |= FAULT_FLAG_WRITE;
if (*flags & FOLL_REMOTE)
@@ -1181,8 +1181,6 @@ static long __get_user_pages(struct mm_struct *mm,
case -ENOMEM:
case -EHWPOISON:
goto out;
- case -ENOENT:
- goto next_page;
}
BUG();
} else if (PTR_ERR(page) == -EEXIST) {
@@ -1823,6 +1821,10 @@ static long __gup_longterm_locked(struct mm_struct *mm,
static bool is_valid_gup_flags(unsigned int gup_flags)
{
+ /* FOLL_MLOCK must be set with FOLL_POPULATE */
+ if (WARN_ON_ONCE((gup_flags & (FOLL_POPULATE | FOLL_MLOCK)) ==
FOLL_MLOCK))
+ return false;
+
/*
* FOLL_PIN must only be set internally by the pin_user_pages*() APIs,
* never directly by the caller, so enforce that with an assertion:
> No-tree user does this so that bug is what ever
> crap out of tree code you're using.
populate_vma_page_range() may trigger this bug, but I haven't verified it yet.
populate_vma_page_range
gup_flags = FOLL_TOUCH | FOLL_POPULATE | FOLL_MLOCK;
if (vma->vm_flags & VM_LOCKONFAULT)
gup_flags &= ~FOLL_POPULATE; // FOLL_MLOCK without FOLL_POPULATE then.
__get_user_pages(..., gup_flags, ...);
--
Thanks
Yafang
next prev parent reply other threads:[~2021-08-30 9:13 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-29 1:19 [PATCH] mm, gpu: fix error when FOLL_MLOCK an unpresent page Yafang Shao
2021-08-30 7:00 ` Christoph Hellwig
2021-08-30 9:12 ` Yafang Shao [this message]
2021-08-30 10:08 ` Christoph Hellwig
2021-08-30 10:55 ` Yafang Shao
2021-08-31 1:36 ` Kirill A. Shutemov
2021-08-31 12:32 ` Yafang Shao
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='CALOAHbBocrx-hTe57Rb=ipOQaN78+syqeb1tbuj72Xak+9wUZw@mail.gmail.com' \
--to=laoar.shao@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=hch@infradead.org \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-mm@kvack.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).