hw/cxl: Two media operations related fixes.
cover hw/cxl: Two media operations related fixes.
Peter reported both of these last cycle. Both are related to sanity checking parameters.
Jonathan Cameron (2):
hw/cxl: Check for overflow on santize media as both base and offset
64bit.
hw/cxl: Take into account how many media operations are requested for
param check
hw/cxl/cxl-mailbox-utils.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
Signature · 1 line
2.48.1
PATCH 1/2 hw/cxl: Check for overflow on santize media as both base and offset 64bit.
The both the size and base of a media sanitize operation are both provided by the VM, an overflow is possible which may result in checks on valid range passing when they should not. Close that by checking for overflow on the addition.
--- hw/cxl/cxl-mailbox-utils.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)
hw/cxl/cxl-mailbox-utils.c +7 -2
@@ -2411,7 +2411,7 @@ static uint64_t get_dc_size(CXLType3Dev *ct3d, MemoryRegion **dc_mr)static int validate_dpa_addr (CXLType3Dev *ct3d,uint64_t dpa_addr,size_t length) { -uint64_t vmr_size, pmr_size, dc_size; +uint64_t vmr_size, pmr_size, dc_size, dpa_end; if ((dpa_addr % CXL_CACHE_LINE_SIZE) || (length % CXL_CACHE_LINE_SIZE) ||
@@ -2423,7 +2423,12 @@ static int validate_dpa_addr(CXLType3Dev *ct3d, uint64_t dpa_addr, pmr_size =get_pmr_size (ct3d,NULL ); dc_size =get_dc_size (ct3d,NULL ); -if (dpa_addr + length > vmr_size + pmr_size + dc_size) { +/* sanitize 64 bit values coming from guest */ +if (uadd64_overflow (dpa_addr, length, &dpa_end)) { +return -EINVAL ; + } + +if (dpa_end > vmr_size + pmr_size + dc_size) {return -EINVAL ; }
Signature · 1 line
2.48.1
PATCH 2/2 hw/cxl: Take into account how many media operations are requested for param check
In reply to Jonathan CameronWhilst the spec doesn't speak to it directly my assumption is that a request for more operations than exist should result in an invalid input error return.
--- hw/cxl/cxl-mailbox-utils.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/cxl/cxl-mailbox-utils.c +1 -1
@@ -2547,7 +2547,7 @@ static CXLRetCode media_operations_discovery(uint8_t *payload_in,* sub class command. */ if (media_op_in_disc_pl->dpa_range_count || - start_index >ARRAY_SIZE (media_op_matrix)) { + start_index + num_ops >ARRAY_SIZE (media_op_matrix)) {return CXL_MBOX_INVALID_INPUT ; }
Signature · 1 line
2.48.1
Quoted 2 lines · On Fri, 2 Jan 2026 15:47:29 +0000 Jonathan Cameron…
Peter reported both of these last cycle. Both are related to sanity checking parameters.
Hi Michael
Just a reminder to take a look at and if possible queue these up.
I see I didn't +CC Peter. Sorry about that Peter. If you could sanity check as they are fixes for bugs you reported a while back.
Thanks,
Jonathan
Quoted 10 lines
Jonathan Cameron (2):
hw/cxl: Check for overflow on santize media as both base and offset64bit.hw/cxl: Take into account how many media operations are requested for
param checkhw/cxl/cxl-mailbox-utils.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
Quoted 11 lines · Jonathan Cameron via
Peter reported both of these last cycle. Both are related to sanity checking parameters.
Jonathan Cameron (2):
hw/cxl: Check for overflow on santize media as both base and offset64bit.hw/cxl: Take into account how many media operations are requested for
param checkhw/cxl/cxl-mailbox-utils.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
Is this one worth picking up for qemu-stable (10.1.x & 10.2.x)?
Thanks,
/mjt
Quoted 10 lines · On Fri, 6 Feb 2026 01:11:55 +0300 Michael Tokarev <[email protected]> wrote:
Quoted 9 lines · Jonathan Cameron via
Peter reported both of these last cycle. Both are related to sanity checking parameters.
Jonathan Cameron (2):
hw/cxl: Check for overflow on santize media as both base and offset64bit.hw/cxl: Take into account how many media operations are requested for
param check
Quoted 5 lines
Quoted 2 lines
hw/cxl/cxl-mailbox-utils.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)Is this one worth picking up for qemu-stable (10.1.x & 10.2.x)?
They are fairly obscure features but if it applies cleanly it seems sensible to backport.
Thanks
Jonathan
Quoted 3 lines
Thanks,
/mjt