privcmd_vm_ops defines .close (privcmd_close), but neither .may_split
nor .open. When userspace does a partial munmap() on a privcmd mapping,
the kernel splits the VMA via __split_vma(). Since may_split is NULL,
the split is allowed. vm_area_dup() copies vm_private_data (a pages
array allocated in alloc_empty_pages()) into the new VMA without any
fixup, because there is no .open callback.
Both VMAs now point to the same pages array. When the unmapped portion
is closed, privcmd_close() calls:
- xen_unmap_domain_gfn_range()
- xen_free_unpopulated_pages()
- kvfree(pages)
The surviving VMA still holds the dangling pointer. When it is later
destroyed, the same sequence runs again, which leads to a double free.
Fix this issue by adding a .may_split callback denying the VMA split.
This is XSA-487 / CVE-2026-31787
Fixes: d71f513985 ("xen: privcmd: support autotranslated physmap guests.")
Reported-by: Atharva Vartak <atharva.a.vartak@gmail.com>
Suggested-by: Atharva Vartak <atharva.a.vartak@gmail.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Commit 453b8fb68f ("xen/privcmd: restrict usage in
unprivileged domU") added a xenstore notifier to defer setting the
restriction target until Xenstore is ready.
XEN_PRIVCMD can be built as a module, but privcmd_exit() leaves that
notifier behind. Balance the notifier lifecycle by unregistering it on
module exit.
This is harmless even if xenstore was already ready at registration
time and the notifier was never queued on the chain.
Fixes: 453b8fb68f ("xen/privcmd: restrict usage in unprivileged domU")
Signed-off-by: GuoHan Zhao <zhaoguohan@kylinos.cn>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20260325120246.252899-1-zhaoguohan@kylinos.cn>
When running in an unprivileged domU under Xen, the privcmd driver
is restricted to allow only hypercalls against a target domain, for
which the current domU is acting as a device model.
Add a boot parameter "unrestricted" to allow all hypercalls (the
hypervisor will still refuse destructive hypercalls affecting other
guests).
Make this new parameter effective only in case the domU wasn't started
using secure boot, as otherwise hypercalls targeting the domU itself
might result in violating the secure boot functionality.
This is achieved by adding another lockdown reason, which can be
tested to not being set when applying the "unrestricted" option.
This is part of XSA-482
Signed-off-by: Juergen Gross <jgross@suse.com>
---
V2:
- new patch
The Xen privcmd driver allows to issue arbitrary hypercalls from
user space processes. This is normally no problem, as access is
usually limited to root and the hypervisor will deny any hypercalls
affecting other domains.
In case the guest is booted using secure boot, however, the privcmd
driver would be enabling a root user process to modify e.g. kernel
memory contents, thus breaking the secure boot feature.
The only known case where an unprivileged domU is really needing to
use the privcmd driver is the case when it is acting as the device
model for another guest. In this case all hypercalls issued via the
privcmd driver will target that other guest.
Fortunately the privcmd driver can already be locked down to allow
only hypercalls targeting a specific domain, but this mode can be
activated from user land only today.
The target domain can be obtained from Xenstore, so when not running
in dom0 restrict the privcmd driver to that target domain from the
beginning, resolving the potential problem of breaking secure boot.
This is XSA-482
Reported-by: Teddy Astie <teddy.astie@vates.tech>
Fixes: 1c5de1939c ("xen: add privcmd driver")
Signed-off-by: Juergen Gross <jgross@suse.com>
---
V2:
- defer reading from Xenstore if Xenstore isn't ready yet (Jan Beulich)
- wait in open() if target domain isn't known yet
- issue message in case no target domain found (Jan Beulich)
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
Currently if a user enqueue a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.
This lack of consistentcy cannot be addressed without refactoring the API.
alloc_workqueue() treats all queues as per-CPU by default, while unbound
workqueues must opt-in via WQ_UNBOUND.
This default is suboptimal: most workloads benefit from unbound queues,
allowing the scheduler to place worker threads where they’re needed and
reducing noise when CPUs are isolated.
This continues the effort to refactor workqueue APIs, which began with
the introduction of new workqueues and a new alloc_workqueue flag in:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
This change adds a new WQ_PERCPU flag to explicitly request alloc_workqueue()
to be per-cpu when WQ_UNBOUND has not been specified.
With the introduction of the WQ_PERCPU flag (equivalent to !WQ_UNBOUND),
any alloc_workqueue() caller that doesn’t explicitly specify WQ_UNBOUND
must now use WQ_PERCPU.
Once migration is complete, WQ_UNBOUND can be removed and unbound will
become the implicit default.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20251106155831.306248-3-marco.crivellari@suse.com>
Instead of testing the XENFEAT_auto_translated_physmap feature, just
use !xen_pv_domain() which is equivalent.
This has the advantage that a kernel not built with CONFIG_XEN_PV
will be smaller due to dead code elimination.
Reviewed-by: Jason Andryuk <jason.andryuk@amd.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20250826145608.10352-3-jgross@suse.com>
Pull 'struct fd' class updates from Al Viro:
"The bulk of struct fd memory safety stuff
Making sure that struct fd instances are destroyed in the same scope
where they'd been created, getting rid of reassignments and passing
them by reference, converting to CLASS(fd{,_pos,_raw}).
We are getting very close to having the memory safety of that stuff
trivial to verify"
* tag 'pull-fd' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (28 commits)
deal with the last remaing boolean uses of fd_file()
css_set_fork(): switch to CLASS(fd_raw, ...)
memcg_write_event_control(): switch to CLASS(fd)
assorted variants of irqfd setup: convert to CLASS(fd)
do_pollfd(): convert to CLASS(fd)
convert do_select()
convert vfs_dedupe_file_range().
convert cifs_ioctl_copychunk()
convert media_request_get_by_fd()
convert spu_run(2)
switch spufs_calls_{get,put}() to CLASS() use
convert cachestat(2)
convert do_preadv()/do_pwritev()
fdget(), more trivial conversions
fdget(), trivial conversions
privcmd_ioeventfd_assign(): don't open-code eventfd_ctx_fdget()
o2hb_region_dev_store(): avoid goto around fdget()/fdput()
introduce "fd_pos" class, convert fdget_pos() users to it.
fdget_raw() users: switch to CLASS(fd_raw)
convert vmsplice() to CLASS(fd)
...
in all of those failure exits prior to fdget() are plain returns and
the only thing done after fdput() is (on failure exits) a kfree(),
which can be done before fdput() just fine.
NOTE: in acrn_irqfd_assign() 'fail:' failure exit is wrong for
eventfd_ctx_fileget() failure (we only want fdput() there) and once
we stop doing that, it doesn't need to check if eventfd is NULL or
ERR_PTR(...) there.
NOTE: in privcmd we move fdget() up before the allocation - more
to the point, before the copy_from_user() attempt.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Commit 2fae6bb7be ("xen/privcmd: Add new syscall to get gsi from dev")
adds a weak reverse dependency to the config XEN_PRIVCMD definition, that
dependency causes xen-privcmd can't be loaded on domU, because dependent
xen-pciback isn't always be loaded successfully on domU.
To solve above problem, remove that dependency, and do not call
pcistub_get_gsi_from_sbdf() directly, instead add a hook in
drivers/xen/apci.c, xen-pciback register the real call function, then in
privcmd_ioctl_pcidev_get_gsi call that hook.
Fixes: 2fae6bb7be ("xen/privcmd: Add new syscall to get gsi from dev")
Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Message-ID: <20241012084537.1543059-1-Jiqian.Chen@amd.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Pull more xen updates from Juergen Gross:
"A second round of Xen related changes and features:
- a small fix of the xen-pciback driver for a warning issued by
sparse
- support PCI passthrough when using a PVH dom0
- enable loading the kernel in PVH mode at arbitrary addresses,
avoiding conflicts with the memory map when running as a Xen dom0
using the host memory layout"
* tag 'for-linus-6.12-rc1a-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
x86/pvh: Add 64bit relocation page tables
x86/kernel: Move page table macros to header
x86/pvh: Set phys_base when calling xen_prepare_pvh()
x86/pvh: Make PVH entrypoint PIC for x86-64
xen: sync elfnote.h from xen tree
xen/pciback: fix cast to restricted pci_ers_result_t and pci_power_t
xen/privcmd: Add new syscall to get gsi from dev
xen/pvh: Setup gsi for passthrough device
xen/pci: Add a function to reset device for xen
On PVH dom0, when passthrough a device to domU, QEMU and xl tools
want to use gsi number to do pirq mapping, see QEMU code
xen_pt_realize->xc_physdev_map_pirq, and xl code
pci_add_dm_done->xc_physdev_map_pirq, but in current codes, the gsi
number is got from file /sys/bus/pci/devices/<sbdf>/irq, that is
wrong, because irq is not equal with gsi, they are in different
spaces, so pirq mapping fails.
And in current linux codes, there is no method to get gsi
for userspace.
For above purpose, record gsi of pcistub devices when init
pcistub and add a new syscall into privcmd to let userspace
can get gsi when they have a need.
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Message-ID: <20240924061437.2636766-4-Jiqian.Chen@amd.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
For any changes of struct fd representation we need to
turn existing accesses to fields into calls of wrappers.
Accesses to struct fd::flags are very few (3 in linux/file.h,
1 in net/socket.c, 3 in fs/overlayfs/file.c and 3 more in
explicit initializers).
Those can be dealt with in the commit converting to
new layout; accesses to struct fd::file are too many for that.
This commit converts (almost) all of f.file to
fd_file(f). It's not entirely mechanical ('file' is used as
a member name more than just in struct fd) and it does not
even attempt to distinguish the uses in pointer context from
those in boolean context; the latter will be eventually turned
into a separate helper (fd_empty()).
NOTE: mass conversion to fd_empty(), tempting as it
might be, is a bad idea; better do that piecewise in commit
that convert from fdget...() to CLASS(...).
[conflicts in fs/fhandle.c, kernel/bpf/syscall.c, mm/memcontrol.c
caught by git; fs/stat.c one got caught by git grep]
[fs/xattr.c conflict]
Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Nothing prevents simultaneous ioctl calls to privcmd_irqfd_assign() and
privcmd_irqfd_deassign(). If that happens, it is possible that a kirqfd
created and added to the irqfds_list by privcmd_irqfd_assign() may get
removed by another thread executing privcmd_irqfd_deassign(), while the
former is still using it after dropping the locks.
This can lead to a situation where an already freed kirqfd instance may
be accessed and cause kernel oops.
Use SRCU locking to prevent the same, as is done for the KVM
implementation for irqfds.
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/9e884af1f1f842eacbb7afc5672c8feb4dea7f3f.1718703669.git.viresh.kumar@linaro.org
Signed-off-by: Juergen Gross <jgross@suse.com>
With ARCH=x86, make allmodconfig && make W=1 C=1 reports:
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/xen/xen-pciback/xen-pciback.o
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/xen/xen-evtchn.o
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/xen/xen-privcmd.o
Add the missing invocations of the MODULE_DESCRIPTION() macro.
Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20240611-md-drivers-xen-v1-1-1eb677364ca6@quicinc.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Pull misc vfs updates from Christian Brauner:
"This contains the usual miscellaneous features, cleanups, and fixes
for vfs and individual fses.
Features:
- Add Jan Kara as VFS reviewer
- Show correct device and inode numbers in proc/<pid>/maps for vma
files on stacked filesystems. This is now easily doable thanks to
the backing file work from the last cycles. This comes with
selftests
Cleanups:
- Remove a redundant might_sleep() from wait_on_inode()
- Initialize pointer with NULL, not 0
- Clarify comment on access_override_creds()
- Rework and simplify eventfd_signal() and eventfd_signal_mask()
helpers
- Process aio completions in batches to avoid needless wakeups
- Completely decouple struct mnt_idmap from namespaces. We now only
keep the actual idmapping around and don't stash references to
namespaces
- Reformat maintainer entries to indicate that a given subsystem
belongs to fs/
- Simplify fput() for files that were never opened
- Get rid of various pointless file helpers
- Rename various file helpers
- Rename struct file members after SLAB_TYPESAFE_BY_RCU switch from
last cycle
- Make relatime_need_update() return bool
- Use GFP_KERNEL instead of GFP_USER when allocating superblocks
- Replace deprecated ida_simple_*() calls with their current ida_*()
counterparts
Fixes:
- Fix comments on user namespace id mapping helpers. They aren't
kernel doc comments so they shouldn't be using /**
- s/Retuns/Returns/g in various places
- Add missing parameter documentation on can_move_mount_beneath()
- Rename i_mapping->private_data to i_mapping->i_private_data
- Fix a false-positive lockdep warning in pipe_write() for watch
queues
- Improve __fget_files_rcu() code generation to improve performance
- Only notify writer that pipe resizing has finished after setting
pipe->max_usage otherwise writers are never notified that the pipe
has been resized and hang
- Fix some kernel docs in hfsplus
- s/passs/pass/g in various places
- Fix kernel docs in ntfs
- Fix kcalloc() arguments order reported by gcc 14
- Fix uninitialized value in reiserfs"
* tag 'vfs-6.8.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (36 commits)
reiserfs: fix uninit-value in comp_keys
watch_queue: fix kcalloc() arguments order
ntfs: dir.c: fix kernel-doc function parameter warnings
fs: fix doc comment typo fs tree wide
selftests/overlayfs: verify device and inode numbers in /proc/pid/maps
fs/proc: show correct device and inode numbers in /proc/pid/maps
eventfd: Remove usage of the deprecated ida_simple_xx() API
fs: super: use GFP_KERNEL instead of GFP_USER for super block allocation
fs/hfsplus: wrapper.c: fix kernel-doc warnings
fs: add Jan Kara as reviewer
fs/inode: Make relatime_need_update return bool
pipe: wakeup wr_wait after setting max_usage
file: remove __receive_fd()
file: stop exposing receive_fd_user()
fs: replace f_rcuhead with f_task_work
file: remove pointless wrapper
file: s/close_fd_get_file()/file_close_fd()/g
Improve __fget_files_rcu() code generation (and thus __fget_light())
file: massage cleanup of files that failed to open
fs/pipe: Fix lockdep false-positive in watchqueue pipe_write()
...
Fake flexible arrays (zero-length and one-element arrays) are deprecated,
and should be replaced by flexible-array members. So, replace
zero-length array with a flexible-array member in `struct
privcmd_kernel_ioreq`.
Also annotate array `ports` with `__counted_by()` to prepare for the
coming implementation by GCC and Clang of the `__counted_by` attribute.
Flexible array members annotated with `__counted_by` can have their
accesses bounds-checked at run-time via `CONFIG_UBSAN_BOUNDS` (for array
indexing) and `CONFIG_FORTIFY_SOURCE` (for strcpy/memcpy-family functions).
This fixes multiple -Warray-bounds warnings:
drivers/xen/privcmd.c:1239:30: warning: array subscript i is outside array bounds of 'struct ioreq_port[0]' [-Warray-bounds=]
drivers/xen/privcmd.c:1240:30: warning: array subscript i is outside array bounds of 'struct ioreq_port[0]' [-Warray-bounds=]
drivers/xen/privcmd.c:1241:30: warning: array subscript i is outside array bounds of 'struct ioreq_port[0]' [-Warray-bounds=]
drivers/xen/privcmd.c:1245:33: warning: array subscript i is outside array bounds of 'struct ioreq_port[0]' [-Warray-bounds=]
drivers/xen/privcmd.c:1258:67: warning: array subscript i is outside array bounds of 'struct ioreq_port[0]' [-Warray-bounds=]
This results in no differences in binary output.
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/ZVZlg3tPMPCRdteh@work
Signed-off-by: Juergen Gross <jgross@suse.com>
Virtio guests send VIRTIO_MMIO_QUEUE_NOTIFY notification when they need
to notify the backend of an update to the status of the virtqueue. The
backend or another entity, polls the MMIO address for updates to know
when the notification is sent.
It works well if the backend does this polling by itself. But as we move
towards generic backend implementations, we end up implementing this in
a separate user-space program.
Generally, the Virtio backends are implemented to work with the Eventfd
based mechanism. In order to make such backends work with Xen, another
software layer needs to do the polling and send an event via eventfd to
the backend once the notification from guest is received. This results
in an extra context switch.
This is not a new problem in Linux though. It is present with other
hypervisors like KVM, etc. as well. The generic solution implemented in
the kernel for them is to provide an IOCTL call to pass the address to
poll and eventfd, which lets the kernel take care of polling and raise
an event on the eventfd, instead of handling this in user space (which
involves an extra context switch).
This patch adds similar support for xen.
Inspired by existing implementations for KVM, etc..
This also copies ioreq.h header file (only struct ioreq and related
macros) from Xen's source tree (Top commit 5d84f07fe6bf ("xen/pci: drop
remaining uses of bool_t")).
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/b20d83efba6453037d0c099912813c79c81f7714.1697439990.git.viresh.kumar@linaro.org
Signed-off-by: Juergen Gross <jgross@suse.com>
Xen provides support for injecting interrupts to the guests via the
HYPERVISOR_dm_op() hypercall. The same is used by the Virtio based
device backend implementations, in an inefficient manner currently.
Generally, the Virtio backends are implemented to work with the Eventfd
based mechanism. In order to make such backends work with Xen, another
software layer needs to poll the Eventfds and raise an interrupt to the
guest using the Xen based mechanism. This results in an extra context
switch.
This is not a new problem in Linux though. It is present with other
hypervisors like KVM, etc. as well. The generic solution implemented in
the kernel for them is to provide an IOCTL call to pass the interrupt
details and eventfd, which lets the kernel take care of polling the
eventfd and raising of the interrupt, instead of handling this in user
space (which involves an extra context switch).
This patch adds support to inject a specific interrupt to guest using
the eventfd mechanism, by preventing the extra context switch.
Inspired by existing implementations for KVM, etc..
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/8e724ac1f50c2bc1eb8da9b3ff6166f1372570aa.1692697321.git.viresh.kumar@linaro.org
Signed-off-by: Juergen Gross <jgross@suse.com>
Juergen Gross
GuoHan ZhaoandJuergen Gross
Linus Torvalds
Kees Cook
Marco CrivellariandJuergen Gross
Al Viro
Jiqian ChenandJuergen Gross
Viresh KumarandJuergen Gross
Jeff JohnsonandJuergen Gross
Markus ElfringandJuergen Gross
Christian Brauner
Gustavo A. R. SilvaandJuergen Gross