Since pre-git times, is_mergeable_vma() returns false for a vma with
vm_ops->close, so that no owner assumptions are violated in case the vma
is removed as part of the merge.
This check is currently very conservative and can prevent merging even
situations where vma can't be removed, such as simple expansion of
previous vma, as evidenced by commit d014cd7c1c ("mm, mremap: fix
mremap() expanding for vma's with vm_ops->close()")
In order to allow more merging when appropriate and simplify the code that
was made more complex by commit d014cd7c1c, start distinguishing cases
where the vma can be really removed, and allow merging with vm_ops->close
otherwise.
As a first step, add a may_remove_vma parameter to is_mergeable_vma().
can_vma_merge_before() sets it to true, because when called from
vma_merge(), a removal of the vma is possible.
In can_vma_merge_after(), pass the parameter as false, because no
removal can occur in each of its callers:
- vma_merge() calls it on the 'prev' vma, which is never removed
- mmap_region() and do_brk_flags() call it to determine if it can expand
a vma, which is not removed
As a result, vma's with vm_ops->close may now merge with compatible ranges
in more situations than previously. We can also revert commit
d014cd7c1c as the next step to simplify mremap code again.
[vbabka@suse.cz: adjust comment as suggested by Lorenzo]
Link: https://lkml.kernel.org/r/74f2ea6c-f1a9-6dd7-260c-25e660f42379@suse.cz
Link: https://lkml.kernel.org/r/20230309111258.24079-10-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Cc: Lorenzo Stoakes <lstoakes@gmail.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Case 1 is now shown in the comment as next vma being merged with prev, so
use 'next' instead of 'mid'. In case 1 they both point to the same vma.
As a consequence, in case 6, the dup_anon_vma() is now tried first on
'next' and then on 'mid', before it was the opposite order. This is not a
functional change, as those two vma's cannnot have a different anon_vma,
as that would have prevented the merging in the first place.
Link: https://lkml.kernel.org/r/20230309111258.24079-4-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Lorenzo Stoakes <lstoakes@gmail.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In case 3 we we use 'next' for everything but vma_pgoff. So use 'next'
for that as well, instead of 'mid', for consistency. Then in case 8 we
have to use 'mid' explicitly, which should also make the intent more
obvious.
Adjust the diagram for cases 1-3 in the comment to match the code - we are
using 'next' for case 3 so mark the range with XXXX instead of NNNN. For
case 2 that's a no-op as the code doesn't touch 'next' or 'mid'. For case
1 it's now wrong but that will be fixed next.
No functional change.
Link: https://lkml.kernel.org/r/20230309111258.24079-3-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Lorenzo Stoakes <lstoakes@gmail.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "cleanup vma_merge() and improve mergeability tests".
My initial goal here was to try making the check for vm_ops->close in
is_mergeable_vma() only be applied for vma's that would be truly removed
as part of the merge (see Patch 9). This would then allow reverting the
quick fix d014cd7c1c ("mm, mremap: fix mremap() expanding for vma's with
vm_ops->close()"). This was successful enough to allow the revert (Patch
10). Checks using can_vma_merge_before() are still pessimistic about
possible vma removal, and making them precise would probably complicate
the vma_merge() code too much.
Liam's 6.3-rc1 simplification of vma_merge() and removal of __vma_adjust()
was very much helpful in understanding the vma_merge() implementation and
especially when vma removals can happen, which is now very obvious. While
studing the code, I've found ways to make it hopefully even more easy to
follow, so that's the patches 1-8. That made me also notice a bug that's
now already fixed in 6.3-rc1.
This patch (of 10):
In the merging preparation part of vma_merge(), some vma pointer variables
are assigned for later execution of the merge, but also read from in the
block itself. The code is easier follow and check against the cases
diagram in the comment if the code reads only from the "primary" vma
variables prev, mid, next instead. No functional change.
Link: https://lkml.kernel.org/r/20230309111258.24079-1-vbabka@suse.cz
Link: https://lkml.kernel.org/r/20230309111258.24079-2-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Lorenzo Stoakes <lstoakes@gmail.com>]
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Use the maple tree in RCU mode for VMA tracking.
The maple tree tracks the stack and is able to update the pivot
(lower/upper boundary) in-place to allow the page fault handler to write
to the tree while holding just the mmap read lock. This is safe as the
writes to the stack have a guard VMA which ensures there will always be a
NULL in the direction of the growth and thus will only update a pivot.
It is possible, but not recommended, to have VMAs that grow up/down
without guard VMAs. syzbot has constructed a testcase which sets up a VMA
to grow and consume the empty space. Overwriting the entire NULL entry
causes the tree to be altered in a way that is not safe for concurrent
readers; the readers may see a node being rewritten or one that does not
match the maple state they are using.
Enabling RCU mode allows the concurrent readers to see a stable node and
will return the expected result.
[Liam.Howlett@Oracle.com: we don't need to free the nodes with RCU[
Link: https://lore.kernel.org/linux-mm/000000000000b0a65805f663ace6@google.com/
Link: https://lkml.kernel.org/r/20230227173632.3292573-9-surenb@google.com
Fixes: d4af56c5c7 ("mm: start tracking VMAs with maple tree")
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Reported-by: syzbot+8d95422d3537159ca390@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In case 4, we are shrinking 'prev' (PPPP in the comment) and expanding
'mid' (NNNN). So we need to make sure 'mid' clones the anon_vma from
'prev', if it doesn't have any. After commit 0503ea8f5b ("mm/mmap:
remove __vma_adjust()") we can fail to do that due to wrong parameters for
dup_anon_vma(). The call is a no-op because res == next, adjust == mid
and mid == next. Fix it.
Link: https://lkml.kernel.org/r/ad91d62b-37eb-4b73-707a-3c45c9e16256@suse.cz
Fixes: 0503ea8f5b ("mm/mmap: remove __vma_adjust()")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Introduce shrink_vma() which uses the vma_prepare() and vma_complete()
functions to reduce the vma coverage.
Convert shift_arg_pages() to use expand_vma() and the new shrink_vma()
function. Remove support from __vma_adjust() to reduce a vma size since
shift_arg_pages() is the only user that shrinks a VMA in this way.
Link: https://lkml.kernel.org/r/20230120162650.984577-46-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add init_vma_prep() and init_multi_vma_prep() to set up the struct
vma_prepare. This is to abstract the locking when adjusting the VMAs.
Also change __vma_adjust() variable remove_next int in favour of a pointer
to the VMA to remove. Rename next_next to remove2 since this better
reflects its use.
Link: https://lkml.kernel.org/r/20230120162650.984577-43-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Change the vma_adjust() function definition to accept the vma iterator and
pass it through to __vma_adjust().
Update fs/exec to use the new vma_adjust() function parameters.
Update mm/mremap to use the new vma_adjust() function parameters.
Revert the __split_vma() calls back from __vma_adjust() to vma_adjust()
and pass through the vma iterator.
Link: https://lkml.kernel.org/r/20230120162650.984577-37-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In preparation of passing the vma state through split, the pre-allocation
that occurs before the split has to be moved to after. Since the
preallocation would then live right next to the store, just call store
instead of preallocating. This effectively restores the potential error
path of splitting and not munmap'ing which pre-dates the maple tree.
Link: https://lkml.kernel.org/r/20230120162650.984577-12-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "mm: In-kernel support for memory-deny-write-execute (MDWE)",
v2.
The background to this is that systemd has a configuration option called
MemoryDenyWriteExecute [2], implemented as a SECCOMP BPF filter. Its aim
is to prevent a user task from inadvertently creating an executable
mapping that is (or was) writeable. Since such BPF filter is stateless,
it cannot detect mappings that were previously writeable but subsequently
changed to read-only. Therefore the filter simply rejects any
mprotect(PROT_EXEC). The side-effect is that on arm64 with BTI support
(Branch Target Identification), the dynamic loader cannot change an ELF
section from PROT_EXEC to PROT_EXEC|PROT_BTI using mprotect(). For
libraries, it can resort to unmapping and re-mapping but for the main
executable it does not have a file descriptor. The original bug report in
the Red Hat bugzilla - [3] - and subsequent glibc workaround for libraries
- [4].
This series adds in-kernel support for this feature as a prctl
PR_SET_MDWE, that is inherited on fork(). The prctl denies PROT_WRITE |
PROT_EXEC mappings. Like the systemd BPF filter it also denies adding
PROT_EXEC to mappings. However unlike the BPF filter it only denies it if
the mapping didn't previous have PROT_EXEC. This allows to PROT_EXEC ->
PROT_EXEC | PROT_BTI with mprotect(), which is a problem with the BPF
filter.
This patch (of 2):
The aim of such policy is to prevent a user task from creating an
executable mapping that is also writeable.
An example of mmap() returning -EACCESS if the policy is enabled:
mmap(0, size, PROT_READ | PROT_WRITE | PROT_EXEC, flags, 0, 0);
Similarly, mprotect() would return -EACCESS below:
addr = mmap(0, size, PROT_READ | PROT_EXEC, flags, 0, 0);
mprotect(addr, size, PROT_READ | PROT_WRITE | PROT_EXEC);
The BPF filter that systemd MDWE uses is stateless, and disallows
mprotect() with PROT_EXEC completely. This new prctl allows PROT_EXEC to
be enabled if it was already PROT_EXEC, which allows the following case:
addr = mmap(0, size, PROT_READ | PROT_EXEC, flags, 0, 0);
mprotect(addr, size, PROT_READ | PROT_EXEC | PROT_BTI);
where PROT_BTI enables branch tracking identification on arm64.
Link: https://lkml.kernel.org/r/20230119160344.54358-1-joey.gouly@arm.com
Link: https://lkml.kernel.org/r/20230119160344.54358-2-joey.gouly@arm.com
Signed-off-by: Joey Gouly <joey.gouly@arm.com>
Co-developed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Jeremy Linton <jeremy.linton@arm.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lennart Poettering <lennart@poettering.net>
Cc: Mark Brown <broonie@kernel.org>
Cc: nd <nd@arm.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Szabolcs Nagy <szabolcs.nagy@arm.com>
Cc: Topi Miettinen <toiwoton@gmail.com>
Cc: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Cc: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Currently, we don't enable writenotify when enabling userfaultfd-wp on a
shared writable mapping (for now only shmem and hugetlb). The consequence
is that vma->vm_page_prot will still include write permissions, to be set
as default for all PTEs that get remapped (e.g., mprotect(), NUMA hinting,
page migration, ...).
So far, vma->vm_page_prot is assumed to be a safe default, meaning that we
only add permissions (e.g., mkwrite) but not remove permissions (e.g.,
wrprotect). For example, when enabling softdirty tracking, we enable
writenotify. With uffd-wp on shared mappings, that changed. More details
on vma->vm_page_prot semantics were summarized in [1].
This is problematic for uffd-wp: we'd have to manually check for a uffd-wp
PTEs/PMDs and manually write-protect PTEs/PMDs, which is error prone.
Prone to such issues is any code that uses vma->vm_page_prot to set PTE
permissions: primarily pte_modify() and mk_pte().
Instead, let's enable writenotify such that PTEs/PMDs/... will be mapped
write-protected as default and we will only allow selected PTEs that are
definitely safe to be mapped without write-protection (see
can_change_pte_writable()) to be writable. In the future, we might want
to enable write-bit recovery -- e.g., can_change_pte_writable() -- at more
locations, for example, also when removing uffd-wp protection.
This fixes two known cases:
(a) remove_migration_pte() mapping uffd-wp'ed PTEs writable, resulting
in uffd-wp not triggering on write access.
(b) do_numa_page() / do_huge_pmd_numa_page() mapping uffd-wp'ed PTEs/PMDs
writable, resulting in uffd-wp not triggering on write access.
Note that do_numa_page() / do_huge_pmd_numa_page() can be reached even
without NUMA hinting (which currently doesn't seem to be applicable to
shmem), for example, by using uffd-wp with a PROT_WRITE shmem VMA. On
such a VMA, userfaultfd-wp is currently non-functional.
Note that when enabling userfaultfd-wp, there is no need to walk page
tables to enforce the new default protection for the PTEs: we know that
they cannot be uffd-wp'ed yet, because that can only happen after enabling
uffd-wp for the VMA in general.
Also note that this makes mprotect() on ranges with uffd-wp'ed PTEs not
accidentally set the write bit -- which would result in uffd-wp not
triggering on later write access. This commit makes uffd-wp on shmem
behave just like uffd-wp on anonymous memory in that regard, even though,
mixing mprotect with uffd-wp is controversial.
[1] https://lkml.kernel.org/r/92173bad-caa3-6b43-9d1e-9a471fdbc184@redhat.com
Link: https://lkml.kernel.org/r/20221209080912.7968-1-david@redhat.com
Fixes: b1f9e87686 ("mm/uffd: enable write protection for shmem & hugetlbfs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Reported-by: Ives van Hoorne <ives@codesandbox.io>
Debugged-by: Peter Xu <peterx@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
