From: Catalin Marinas <catalin.marinas@arm.com>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>,
Andreas Gruenbacher <agruenba@redhat.com>,
Christoph Hellwig <hch@infradead.org>,
"Darrick J. Wong" <djwong@kernel.org>, Jan Kara <jack@suse.cz>,
Matthew Wilcox <willy@infradead.org>,
cluster-devel <cluster-devel@redhat.com>,
linux-fsdevel <linux-fsdevel@vger.kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
"ocfs2-devel@oss.oracle.com" <ocfs2-devel@oss.oracle.com>,
Josef Bacik <josef@toxicpanda.com>, Will Deacon <will@kernel.org>
Subject: Re: [RFC][arm64] possible infinite loop in btrfs search_ioctl()
Date: Mon, 11 Oct 2021 22:08:14 +0100 [thread overview]
Message-ID: <YWSnvq58jDsDuIik@arm.com> (raw)
In-Reply-To: <CAHk-=wjvQWj7mvdrgTedUW50c2fkdn6Hzxtsk-=ckkMrFoTXjQ@mail.gmail.com>
On Mon, Oct 11, 2021 at 12:15:43PM -0700, Linus Torvalds wrote:
> On Mon, Oct 11, 2021 at 10:38 AM Catalin Marinas
> <catalin.marinas@arm.com> wrote:
> > I cleaned up this patch [1] but I realised it still doesn't solve it.
> > The arm64 __copy_to_user_inatomic(), while ensuring progress if called
> > in a loop, it does not guarantee precise copy to the fault position.
>
> That should be ok., We've always allowed the user copy to return early
> if it does word copies and hits a page crosser that causes a fault.
>
> Any user then has the choice of:
>
> - partial copies are bad
>
> - partial copies are handled and then you retry from the place
> copy_to_user() failed at
>
> and in that second case, the next time around, you'll get the fault
> immediately (or you'll make some more progress - maybe the user copy
> loop did something different just because the length and/or alignment
> was different).
>
> If you get the fault immediately, that's -EFAULT.
>
> And if you make some more progress, it's again up to the caller to
> rinse and repeat.
>
> End result: user copy functions do not have to report errors exactly.
> It is the caller that has to handle the situation.
>
> Most importantly: "exact error or not" doesn't actually _matter_ to
> the caller. If the caller does the right thing for an exact error, it
> will do the right thing for an inexact one too. See above.
Yes, that's my expectation (though fixed fairly recently in the arm64
user copy routines).
> > The copy_to_sk(), after returning an error, starts again from the previous
> > sizeof(sh) boundary rather than from where the __copy_to_user_inatomic()
> > stopped. So it can get stuck attempting to copy the same search header.
>
> That seems to be purely a copy_to_sk() bug.
>
> Or rather, it looks like a bug in the caller. copy_to_sk() itself does
>
> if (copy_to_user_nofault(ubuf + *sk_offset, &sh, sizeof(sh))) {
> ret = 0;
> goto out;
> }
>
> and the comment says
>
> * 0: all items from this leaf copied, continue with next
>
> but that comment is then obviously not actually true in that it's not
> "continue with next" at all.
The comments were correct before commit a48b73eca4ce ("btrfs: fix
potential deadlock in the search ioctl") which introduced the
potentially infinite loop.
Something like this would make the comments match (I think):
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index cc61813213d8..1debf6a124e8 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -2161,7 +2161,7 @@ static noinline int copy_to_sk(struct btrfs_path *path,
* properly this next time through
*/
if (copy_to_user_nofault(ubuf + *sk_offset, &sh, sizeof(sh))) {
- ret = 0;
+ ret = -EFAULT;
goto out;
}
@@ -2175,7 +2175,7 @@ static noinline int copy_to_sk(struct btrfs_path *path,
*/
if (read_extent_buffer_to_user_nofault(leaf, up,
item_off, item_len)) {
- ret = 0;
+ ret = -EFAULT;
*sk_offset -= sizeof(sh);
goto out;
}
@@ -2260,12 +2260,8 @@ static noinline int search_ioctl(struct inode *inode,
key.type = sk->min_type;
key.offset = sk->min_offset;
- while (1) {
- ret = fault_in_pages_writeable(ubuf + sk_offset,
- *buf_size - sk_offset);
- if (ret)
- break;
-
+ ret = fault_in_pages_writeable(ubuf, *buf_size);
+ while (ret == 0) {
ret = btrfs_search_forward(root, &key, path, sk->min_transid);
if (ret != 0) {
if (ret > 0)
@@ -2275,9 +2271,14 @@ static noinline int search_ioctl(struct inode *inode,
ret = copy_to_sk(path, &key, sk, buf_size, ubuf,
&sk_offset, &num_found);
btrfs_release_path(path);
- if (ret)
- break;
+ /*
+ * Fault in copy_to_sk(), attempt to bring the page in after
+ * releasing the locks and retry.
+ */
+ if (ret == -EFAULT)
+ ret = fault_in_pages_writeable(ubuf + sk_offset,
+ sizeof(struct btrfs_ioctl_search_header));
}
if (ret > 0)
ret = 0;
> > An ugly fix is to fall back to byte by byte copying so that we can
> > attempt the actual fault address in fault_in_pages_writeable().
>
> No, changing the user copy machinery is wrong. The caller really has
> to do the right thing with partial results.
>
> And I think we need to make fault_in_pages_writeable() match the
> actual faulting cases - maybe remote the "pages" part of the name?
Ah, good point. Without removing "pages" from the name (too late over
here to grep the kernel), something like below:
diff --git a/arch/arm64/include/asm/page-def.h b/arch/arm64/include/asm/page-def.h
index 2403f7b4cdbf..3768ac4a6610 100644
--- a/arch/arm64/include/asm/page-def.h
+++ b/arch/arm64/include/asm/page-def.h
@@ -15,4 +15,9 @@
#define PAGE_SIZE (_AC(1, UL) << PAGE_SHIFT)
#define PAGE_MASK (~(PAGE_SIZE-1))
+#ifdef CONFIG_ARM64_MTE
+#define FAULT_GRANULE_SIZE (16)
+#define FAULT_GRANULE_MASK (~(FAULT_GRANULE_SIZE-1))
+#endif
+
#endif /* __ASM_PAGE_DEF_H */
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 62db6b0176b9..7aef732e4fa7 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -16,6 +16,11 @@
#include <linux/hardirq.h> /* for in_interrupt() */
#include <linux/hugetlb_inline.h>
+#ifndef FAULT_GRANULE_SIZE
+#define FAULT_GRANULE_SIZE PAGE_SIZE
+#define FAULT_GRANULE_MASK PAGE_MASK
+#endif
+
struct pagevec;
static inline bool mapping_empty(struct address_space *mapping)
@@ -751,12 +756,12 @@ static inline int fault_in_pages_writeable(char __user *uaddr, size_t size)
do {
if (unlikely(__put_user(0, uaddr) != 0))
return -EFAULT;
- uaddr += PAGE_SIZE;
+ uaddr += FAULT_GRANULE_SIZE;
} while (uaddr <= end);
- /* Check whether the range spilled into the next page. */
- if (((unsigned long)uaddr & PAGE_MASK) ==
- ((unsigned long)end & PAGE_MASK))
+ /* Check whether the range spilled into the next granule. */
+ if (((unsigned long)uaddr & FAULT_GRANULE_MASK) ==
+ ((unsigned long)end & FAULT_GRANULE_MASK))
return __put_user(0, end);
return 0;
If this looks in the right direction, I'll do some proper patches
tomorrow.
--
Catalin
next prev parent reply other threads:[~2021-10-11 21:08 UTC|newest]
Thread overview: 99+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-27 16:49 [PATCH v7 00/19] gfs2: Fix mmap + page fault deadlocks Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 01/19] iov_iter: Fix iov_iter_get_pages{,_alloc} page fault return value Andreas Gruenbacher
2021-09-09 11:09 ` Christoph Hellwig
2021-08-27 16:49 ` [PATCH v7 02/19] powerpc/kvm: Fix kvm_use_magic_page Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 03/19] gup: Turn fault_in_pages_{readable,writeable} into fault_in_{readable,writeable} Andreas Gruenbacher
2021-08-27 19:08 ` Al Viro
2021-09-03 14:56 ` Filipe Manana
2021-09-28 15:02 ` Andreas Gruenbacher
2021-09-28 16:37 ` Matthew Wilcox
2021-09-28 20:41 ` Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 04/19] iov_iter: Turn iov_iter_fault_in_readable into fault_in_iov_iter_readable Andreas Gruenbacher
2021-08-27 18:53 ` Al Viro
2021-08-27 18:57 ` Linus Torvalds
2021-08-27 19:16 ` Al Viro
2021-08-27 20:56 ` Kari Argillander
2021-08-28 17:13 ` Linus Torvalds
2021-08-27 16:49 ` [PATCH v7 05/19] iov_iter: Introduce fault_in_iov_iter_writeable Andreas Gruenbacher
2021-08-27 18:49 ` Al Viro
2021-08-27 19:05 ` Linus Torvalds
2021-08-27 19:23 ` Al Viro
2021-08-27 19:33 ` Linus Torvalds
2021-08-27 19:37 ` Al Viro
2021-08-27 21:48 ` Al Viro
2021-08-27 21:57 ` Al Viro
2021-08-27 23:22 ` Luck, Tony
2021-08-28 2:20 ` Luck, Tony
2021-08-28 21:47 ` Thomas Gleixner
2021-08-28 22:04 ` Al Viro
2021-08-28 22:11 ` Al Viro
2021-08-28 22:19 ` Al Viro
2021-08-28 22:51 ` Al Viro
2021-08-29 18:44 ` Thomas Gleixner
2021-08-29 19:46 ` Al Viro
2021-08-29 19:51 ` Thomas Gleixner
2021-08-28 22:20 ` Tony Luck
2021-08-29 1:40 ` Matthew Wilcox
2021-08-30 15:41 ` Luck, Tony
2021-08-28 22:23 ` Thomas Gleixner
2021-08-28 19:28 ` [RFC][arm64] possible infinite loop in btrfs search_ioctl() Al Viro
2021-08-31 13:54 ` Catalin Marinas
2021-08-31 15:28 ` Al Viro
2021-08-31 16:01 ` Catalin Marinas
2021-10-11 17:37 ` Catalin Marinas
2021-10-11 19:15 ` Linus Torvalds
2021-10-11 21:08 ` Catalin Marinas [this message]
2021-10-11 23:59 ` Linus Torvalds
2021-10-12 17:27 ` Catalin Marinas
2021-10-12 17:58 ` Linus Torvalds
2021-10-18 17:13 ` Catalin Marinas
2021-10-21 0:46 ` Andreas Gruenbacher
2021-10-21 10:05 ` Catalin Marinas
2021-10-21 14:42 ` Andreas Gruenbacher
2021-10-21 17:09 ` Catalin Marinas
2021-10-21 18:00 ` Andreas Gruenbacher
2021-10-22 18:41 ` Catalin Marinas
2021-10-25 19:37 ` Andreas Gruenbacher
2021-10-22 2:30 ` Linus Torvalds
2021-10-22 9:34 ` Catalin Marinas
2021-08-29 0:58 ` [PATCH v7 05/19] iov_iter: Introduce fault_in_iov_iter_writeable Al Viro
2021-08-27 16:49 ` [PATCH v7 06/19] gfs2: Add wrapper for iomap_file_buffered_write Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 07/19] gfs2: Clean up function may_grant Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 08/19] gfs2: Eliminate vestigial HIF_FIRST Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 09/19] gfs2: Remove redundant check from gfs2_glock_dq Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 10/19] gfs2: Introduce flag for glock holder auto-demotion Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 11/19] gfs2: Move the inode glock locking to gfs2_file_buffered_write Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 12/19] gfs2: Eliminate ip->i_gh Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 13/19] gfs2: Fix mmap + page fault deadlocks for buffered I/O Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 14/19] iomap: Fix iomap_dio_rw return value for user copies Andreas Gruenbacher
2021-09-03 18:54 ` Darrick J. Wong
2021-09-09 11:17 ` Christoph Hellwig
2021-08-27 16:49 ` [PATCH v7 15/19] iomap: Support partial direct I/O on user copy failures Andreas Gruenbacher
2021-09-03 18:54 ` Darrick J. Wong
2021-09-09 11:20 ` Christoph Hellwig
2021-09-28 15:05 ` Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 16/19] iomap: Add done_before argument to iomap_dio_rw Andreas Gruenbacher
2021-08-27 18:30 ` Darrick J. Wong
2021-08-27 20:15 ` Andreas Gruenbacher
2021-08-27 21:32 ` Darrick J. Wong
2021-08-27 21:49 ` Andreas Grünbacher
2021-08-27 22:35 ` Linus Torvalds
2021-09-03 18:47 ` Darrick J. Wong
2021-09-03 18:53 ` Darrick J. Wong
2021-09-09 11:30 ` Christoph Hellwig
2021-09-09 17:22 ` Linus Torvalds
2021-09-10 7:36 ` Christoph Hellwig
2021-08-27 16:49 ` [PATCH v7 17/19] gup: Introduce FOLL_NOFAULT flag to disable page faults Andreas Gruenbacher
2021-09-09 11:36 ` Christoph Hellwig
2021-09-09 17:17 ` Linus Torvalds
2021-09-10 7:24 ` Christoph Hellwig
2021-08-27 16:49 ` [PATCH v7 18/19] iov_iter: Introduce nofault " Andreas Gruenbacher
2021-08-27 18:47 ` Al Viro
2021-08-27 19:56 ` Andreas Gruenbacher
2021-08-27 16:49 ` [PATCH v7 19/19] gfs2: Fix mmap + page fault deadlocks for direct I/O Andreas Gruenbacher
2021-08-27 17:16 ` [PATCH v7 00/19] gfs2: Fix mmap + page fault deadlocks Linus Torvalds
2021-09-01 19:52 ` Andreas Gruenbacher
2021-09-03 15:52 ` Linus Torvalds
2021-09-03 18:25 ` Al Viro
2021-09-03 18:47 ` Linus Torvalds
2021-09-03 15:07 ` Filipe Manana
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=YWSnvq58jDsDuIik@arm.com \
--to=catalin.marinas@arm.com \
--cc=agruenba@redhat.com \
--cc=cluster-devel@redhat.com \
--cc=djwong@kernel.org \
--cc=hch@infradead.org \
--cc=jack@suse.cz \
--cc=josef@toxicpanda.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ocfs2-devel@oss.oracle.com \
--cc=torvalds@linux-foundation.org \
--cc=viro@zeniv.linux.org.uk \
--cc=will@kernel.org \
--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).