* [PATCH] hw/qxl: move check of slot_id before accessing guest_slots
@ 2023-09-14 9:27 Anastasia Belova
2023-09-14 12:32 ` Michael Tokarev
2023-09-14 13:09 ` Philippe Mathieu-Daudé
0 siblings, 2 replies; 3+ messages in thread
From: Anastasia Belova @ 2023-09-14 9:27 UTC (permalink / raw)
To: Gerd Hoffmann; +Cc: Anastasia Belova, qemu-devel, sdl.qemu
If slot_id >= NUM_MEMSLOTS, buffer overflow is possible.
So the check should be upper than d->guest_slots[slot_id]
where size of d->guest_slots is NUM_MEMSLOTS.
Fixes: e954ea2873 ("qxl: qxl_add_memslot: remove guest trigerrable panics")
Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
---
hw/display/qxl.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/hw/display/qxl.c b/hw/display/qxl.c
index 7bb00d68f5..dc618727c0 100644
--- a/hw/display/qxl.c
+++ b/hw/display/qxl.c
@@ -1309,16 +1309,17 @@ static int qxl_add_memslot(PCIQXLDevice *d, uint32_t slot_id, uint64_t delta,
QXLDevMemSlot memslot;
int i;
- guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
- guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
-
- trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
-
if (slot_id >= NUM_MEMSLOTS) {
qxl_set_guest_bug(d, "%s: slot_id >= NUM_MEMSLOTS %d >= %d", __func__,
slot_id, NUM_MEMSLOTS);
return 1;
}
+
+ guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
+ guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
+
+ trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
+
if (guest_start > guest_end) {
qxl_set_guest_bug(d, "%s: guest_start > guest_end 0x%" PRIx64
" > 0x%" PRIx64, __func__, guest_start, guest_end);
--
2.30.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] hw/qxl: move check of slot_id before accessing guest_slots
2023-09-14 9:27 [PATCH] hw/qxl: move check of slot_id before accessing guest_slots Anastasia Belova
@ 2023-09-14 12:32 ` Michael Tokarev
2023-09-14 13:09 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 3+ messages in thread
From: Michael Tokarev @ 2023-09-14 12:32 UTC (permalink / raw)
To: Anastasia Belova, Gerd Hoffmann; +Cc: qemu-devel, sdl.qemu
14.09.2023 12:27, Anastasia Belova wrote:
> If slot_id >= NUM_MEMSLOTS, buffer overflow is possible.
> So the check should be upper than d->guest_slots[slot_id]
> where size of d->guest_slots is NUM_MEMSLOTS.
>
> Fixes: e954ea2873 ("qxl: qxl_add_memslot: remove guest trigerrable panics")
> Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
Good catch.
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
The change is trivial enough for trivial-patches tree,
picked it up if Gerd don't mind.
/mjt
> hw/display/qxl.c | 11 ++++++-----
> 1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/hw/display/qxl.c b/hw/display/qxl.c
> index 7bb00d68f5..dc618727c0 100644
> --- a/hw/display/qxl.c
> +++ b/hw/display/qxl.c
> @@ -1309,16 +1309,17 @@ static int qxl_add_memslot(PCIQXLDevice *d, uint32_t slot_id, uint64_t delta,
> QXLDevMemSlot memslot;
> int i;
>
> - guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
> - guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
> -
> - trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
> -
> if (slot_id >= NUM_MEMSLOTS) {
> qxl_set_guest_bug(d, "%s: slot_id >= NUM_MEMSLOTS %d >= %d", __func__,
> slot_id, NUM_MEMSLOTS);
> return 1;
> }
> +
> + guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
> + guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
> +
> + trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
> +
> if (guest_start > guest_end) {
> qxl_set_guest_bug(d, "%s: guest_start > guest_end 0x%" PRIx64
> " > 0x%" PRIx64, __func__, guest_start, guest_end);
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] hw/qxl: move check of slot_id before accessing guest_slots
2023-09-14 9:27 [PATCH] hw/qxl: move check of slot_id before accessing guest_slots Anastasia Belova
2023-09-14 12:32 ` Michael Tokarev
@ 2023-09-14 13:09 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 3+ messages in thread
From: Philippe Mathieu-Daudé @ 2023-09-14 13:09 UTC (permalink / raw)
To: Anastasia Belova, Gerd Hoffmann; +Cc: qemu-devel, sdl.qemu, Michael Tokarev
Hi Anastasia,
On 14/9/23 11:27, Anastasia Belova wrote:
> If slot_id >= NUM_MEMSLOTS, buffer overflow is possible.
overflow: unlikely. Do you mean over-read?
Did you found that by code audit? I can't see where this
function get slot_id >= NUM_MEMSLOTS.
This isn't guest triggerable and seems an API invalid use,
so developer error, which we can catch with assertions:
-- >8 --
diff --git a/hw/display/qxl.c b/hw/display/qxl.c
index 7bb00d68f5..443f034168 100644
--- a/hw/display/qxl.c
+++ b/hw/display/qxl.c
@@ -1309,16 +1309,13 @@ static int qxl_add_memslot(PCIQXLDevice *d,
uint32_t slot_id, uint64_t delta,
QXLDevMemSlot memslot;
int i;
+ assert(slot_id < NUM_MEMSLOTS);
+
guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
- if (slot_id >= NUM_MEMSLOTS) {
- qxl_set_guest_bug(d, "%s: slot_id >= NUM_MEMSLOTS %d >= %d",
__func__,
- slot_id, NUM_MEMSLOTS);
- return 1;
- }
if (guest_start > guest_end) {
qxl_set_guest_bug(d, "%s: guest_start > guest_end 0x%" PRIx64
" > 0x%" PRIx64, __func__, guest_start,
guest_end);
---
Note, qxl_add_memslot() return value is ignored...
> So the check should be upper than d->guest_slots[slot_id]
> where size of d->guest_slots is NUM_MEMSLOTS.
>
> Fixes: e954ea2873 ("qxl: qxl_add_memslot: remove guest trigerrable panics")
> Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
> ---
> hw/display/qxl.c | 11 ++++++-----
> 1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/hw/display/qxl.c b/hw/display/qxl.c
> index 7bb00d68f5..dc618727c0 100644
> --- a/hw/display/qxl.c
> +++ b/hw/display/qxl.c
> @@ -1309,16 +1309,17 @@ static int qxl_add_memslot(PCIQXLDevice *d, uint32_t slot_id, uint64_t delta,
> QXLDevMemSlot memslot;
> int i;
>
> - guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
> - guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
> -
> - trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
> -
> if (slot_id >= NUM_MEMSLOTS) {
> qxl_set_guest_bug(d, "%s: slot_id >= NUM_MEMSLOTS %d >= %d", __func__,
> slot_id, NUM_MEMSLOTS);
> return 1;
> }
> +
> + guest_start = le64_to_cpu(d->guest_slots[slot_id].slot.mem_start);
> + guest_end = le64_to_cpu(d->guest_slots[slot_id].slot.mem_end);
> +
> + trace_qxl_memslot_add_guest(d->id, slot_id, guest_start, guest_end);
> +
> if (guest_start > guest_end) {
> qxl_set_guest_bug(d, "%s: guest_start > guest_end 0x%" PRIx64
> " > 0x%" PRIx64, __func__, guest_start, guest_end);
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-09-14 13:10 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-09-14 9:27 [PATCH] hw/qxl: move check of slot_id before accessing guest_slots Anastasia Belova
2023-09-14 12:32 ` Michael Tokarev
2023-09-14 13:09 ` Philippe Mathieu-Daudé
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).