On Tue, Jan 31, 2023 at 8:18 AM Martin Jansa via lists.openembedded.org
<Martin.Jansa=gmail....@lists.openembedded.org> wrote:
> I haven't checked yet in dunfell, but in kirkstone this commit merged last
> week breaks nativesdk-qemu builds.
>
I can confirm it fails the same with this commit from Hitendra in dunfell:
nativesdk-qemu/4.2.0-r0/qemu-4.2.0/hw/display/qxl.c:1508:67: error: 'size'
undeclared (first use in this function); did you mean 'gsize'?
| 1508 | if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset,
size)) {
| |
^~~~
| |
gsize
as well as the same backport from Bhabu in kirkstone (already merged):
https://git.openembedded.org/openembedded-core/commit/?h=kirkstone&id=4cb3874abf4fdeb04337a48a14c765ba9b2269d4
To reproduce this just enable "spice" PACKAGECONFIG in nativesdk-qemu.
Hitendra, Bhabu: please have a look.
Regards,
>
> This chunk:
> +@@ -1486,7 +1505,7 @@ void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL
> pqxl, int group_id)
> + offset = le64_to_cpu(pqxl) & 0xffffffffffff;
> + return (void *)(intptr_t)offset;
> + case MEMSLOT_GROUP_GUEST:
> +- if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset)) {
> ++ if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset, size))
> {
> + return NULL;
> + }
> + ptr = memory_region_get_ram_ptr(qxl->guest_slots[slot].mr);
>
> Uses size parameter which was added in previous commit (which wasn't
> backported):
>
> https://gitlab.com/qemu-project/qemu/-/commit/8efec0ef8bbc1e75a7ebf6e325a35806ece9b39f
>
> So either both commits need to be backported or this one reworked not to
> use undeclared size as it leads to:
>
> | ../qemu-6.2.0/hw/display/qxl.c: In function 'qxl_phys2virt':
> | ../qemu-6.2.0/hw/display/qxl.c:1477:67: error: 'size' undeclared (first
> use in this function); did you mean 'gsize'?
> | 1477 | if (!qxl_get_check_slot_offset(qxl, pqxl, &slot,
> &offset, size)) {
> | |
> ^~~~
> | |
> gsize
>
> I'm surprised this wasn't caught on autobuilder already, maybe it's
> triggered only with extra PACKAGECONFIG options we have enabled:
> PACKAGECONFIG:append:class-nativesdk = " virglrenderer epoxy spice libusb
> usb-redir"
>
> Regards,
>
>
> On Wed, Jan 25, 2023 at 3:42 PM Steve Sakoman <st...@sakoman.com> wrote:
>
>> From: Hitendra Prajapati <hprajap...@mvista.com>
>>
>> Upstream-Status: Backport from
>> https://gitlab.com/qemu-project/qemu/-/commit/6dbbf055148c6f1b7d8a3251a65bd6f3d1e1f622
>>
>> Signed-off-by: Hitendra Prajapati <hprajap...@mvista.com>
>> Signed-off-by: Steve Sakoman <st...@sakoman.com>
>> ---
>> meta/recipes-devtools/qemu/qemu.inc | 1 +
>> .../qemu/qemu/CVE-2022-4144.patch | 103 ++++++++++++++++++
>> 2 files changed, 104 insertions(+)
>> create mode 100644 meta/recipes-devtools/qemu/qemu/CVE-2022-4144.patch
>>
>> diff --git a/meta/recipes-devtools/qemu/qemu.inc
>> b/meta/recipes-devtools/qemu/qemu.inc
>> index fff2c87780..898fa1a8d8 100644
>> --- a/meta/recipes-devtools/qemu/qemu.inc
>> +++ b/meta/recipes-devtools/qemu/qemu.inc
>> @@ -115,6 +115,7 @@ SRC_URI = "
>> https://download.qemu.org/${BPN}-${PV}.tar.xz \
>> file://CVE-2021-3638.patch \
>> file://CVE-2021-20196.patch \
>> file://CVE-2021-3507.patch \
>> + file://CVE-2022-4144.patch \
>> "
>> UPSTREAM_CHECK_REGEX = "qemu-(?P<pver>\d+(\.\d+)+)\.tar"
>>
>> diff --git a/meta/recipes-devtools/qemu/qemu/CVE-2022-4144.patch
>> b/meta/recipes-devtools/qemu/qemu/CVE-2022-4144.patch
>> new file mode 100644
>> index 0000000000..3f0d5fbd5c
>> --- /dev/null
>> +++ b/meta/recipes-devtools/qemu/qemu/CVE-2022-4144.patch
>> @@ -0,0 +1,103 @@
>> +From 6dbbf055148c6f1b7d8a3251a65bd6f3d1e1f622 Mon Sep 17 00:00:00 2001
>> +From: =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= <phi...@linaro.org>
>> +Date: Mon, 28 Nov 2022 21:27:40 +0100
>> +Subject: [PATCH] hw/display/qxl: Avoid buffer overrun in qxl_phys2virt
>> + (CVE-2022-4144)
>> +MIME-Version: 1.0
>> +Content-Type: text/plain; charset=UTF-8
>> +Content-Transfer-Encoding: 8bit
>> +
>> +Have qxl_get_check_slot_offset() return false if the requested
>> +buffer size does not fit within the slot memory region.
>> +
>> +Similarly qxl_phys2virt() now returns NULL in such case, and
>> +qxl_dirty_one_surface() aborts.
>> +
>> +This avoids buffer overrun in the host pointer returned by
>> +memory_region_get_ram_ptr().
>> +
>> +Fixes: CVE-2022-4144 (out-of-bounds read)
>> +Reported-by: Wenxu Yin (@awxylitol)
>> +Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1336
>> +
>> +Signed-off-by: Philippe Mathieu-Daudé <phi...@linaro.org>
>> +Signed-off-by: Stefan Hajnoczi <stefa...@redhat.com>
>> +Message-Id: <20221128202741.4945-5-phi...@linaro.org>
>> +
>> +Upstream-Status: Backport [
>> https://gitlab.com/qemu-project/qemu/-/commit/6dbbf055148c6f1b7d8a3251a65bd6f3d1e1f622
>> ]
>> +CVE: CVE-2022-4144
>> +Comments: Deleted patch hunk in qxl.h,as it contains change
>> +in comments which is not present in current version of qemu.
>> +
>> +Signed-off-by: Hitendra Prajapati <hprajap...@mvista.com>
>> +---
>> + hw/display/qxl.c | 27 +++++++++++++++++++++++----
>> + 1 file changed, 23 insertions(+), 4 deletions(-)
>> +
>> +diff --git a/hw/display/qxl.c b/hw/display/qxl.c
>> +index cd7eb39d..6bc8385b 100644
>> +--- a/hw/display/qxl.c
>> ++++ b/hw/display/qxl.c
>> +@@ -1440,11 +1440,13 @@ static void qxl_reset_surfaces(PCIQXLDevice *d)
>> +
>> + /* can be also called from spice server thread context */
>> + static bool qxl_get_check_slot_offset(PCIQXLDevice *qxl, QXLPHYSICAL
>> pqxl,
>> +- uint32_t *s, uint64_t *o)
>> ++ uint32_t *s, uint64_t *o,
>> ++ size_t size_requested)
>> + {
>> + uint64_t phys = le64_to_cpu(pqxl);
>> + uint32_t slot = (phys >> (64 - 8)) & 0xff;
>> + uint64_t offset = phys & 0xffffffffffff;
>> ++ uint64_t size_available;
>> +
>> + if (slot >= NUM_MEMSLOTS) {
>> + qxl_set_guest_bug(qxl, "slot too large %d >= %d", slot,
>> +@@ -1468,6 +1470,23 @@ static bool
>> qxl_get_check_slot_offset(PCIQXLDevice *qxl, QXLPHYSICAL pqxl,
>> + slot, offset, qxl->guest_slots[slot].size);
>> + return false;
>> + }
>> ++ size_available = memory_region_size(qxl->guest_slots[slot].mr);
>> ++ if (qxl->guest_slots[slot].offset + offset >= size_available) {
>> ++ qxl_set_guest_bug(qxl,
>> ++ "slot %d offset %"PRIu64" > region size
>> %"PRIu64"\n",
>> ++ slot, qxl->guest_slots[slot].offset + offset,
>> ++ size_available);
>> ++ return false;
>> ++ }
>> ++ size_available -= qxl->guest_slots[slot].offset + offset;
>> ++ if (size_requested > size_available) {
>> ++ qxl_set_guest_bug(qxl,
>> ++ "slot %d offset %"PRIu64" size %zu: "
>> ++ "overrun by %"PRIu64" bytes\n",
>> ++ slot, offset, size_requested,
>> ++ size_requested - size_available);
>> ++ return false;
>> ++ }
>> +
>> + *s = slot;
>> + *o = offset;
>> +@@ -1486,7 +1505,7 @@ void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL
>> pqxl, int group_id)
>> + offset = le64_to_cpu(pqxl) & 0xffffffffffff;
>> + return (void *)(intptr_t)offset;
>> + case MEMSLOT_GROUP_GUEST:
>> +- if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset)) {
>> ++ if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset,
>> size)) {
>> + return NULL;
>> + }
>> + ptr = memory_region_get_ram_ptr(qxl->guest_slots[slot].mr);
>> +@@ -1944,9 +1963,9 @@ static void qxl_dirty_one_surface(PCIQXLDevice
>> *qxl, QXLPHYSICAL pqxl,
>> + uint32_t slot;
>> + bool rc;
>> +
>> +- rc = qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset);
>> +- assert(rc == true);
>> + size = (uint64_t)height * abs(stride);
>> ++ rc = qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset, size);
>> ++ assert(rc == true);
>> + trace_qxl_surfaces_dirty(qxl->id, offset, size);
>> + qxl_set_dirty(qxl->guest_slots[slot].mr,
>> + qxl->guest_slots[slot].offset + offset,
>> +--
>> +2.25.1
>> +
>> --
>> 2.25.1
>>
>>
>>
>>
>>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#176508):
https://lists.openembedded.org/g/openembedded-core/message/176508
Mute This Topic: https://lists.openembedded.org/mt/96521255/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
