vfio/listener: Only warn on misalignment for VFIO-owned regions
PATCH 1/1 vfio/listener: Only warn on misalignment for VFIO-owned regions
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the vfio_known_safe_misalignment() whitelist with a vfio_get_vfio_device() ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into a vfio_section_misaligned() helper for readability.
Note: vfio_get_vfio_device() currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.
--- hw/vfio/device.c | 2 +- hw/vfio/listener.c | 33 +++++++++++++++------------------ hw/vfio/trace-events | 2 +- 3 files changed, 17 insertions(+), 20 deletions(-)
hw/vfio/device.c +1 -1
@@ -439,7 +439,7 @@ bool vfio_device_hiod_create_and_realize(VFIODevice *vbasedev,VFIODevice *vfio_get_vfio_device (Object *obj) { -if (object_dynamic_cast (obj,TYPE_VFIO_PCI )) { +if (object_dynamic_cast (obj,TYPE_VFIO_PCI_DEVICE )) {return &VFIO_PCI_DEVICE (obj)->vbasedev; }else {return NULL ;
hw/vfio/listener.c +15 -18
@@ -40,7 +40,6 @@#include "migration/misc.h" #include "migration/qemu-file.h" #include "system/tcg.h" -#include "system/tpm.h" #include "vfio-migration-internal.h" #include "vfio-helpers.h" #include "vfio-listener.h"
@@ -352,20 +351,10 @@ static void vfio_ram_discard_unregister_listener(VFIOContainer *bcontainer,g_free (vrdl); } -static bool vfio_known_safe_misalignment (MemoryRegionSection *section) +static bool vfio_section_misaligned (MemoryRegionSection *section) { -MemoryRegion *mr = section->mr; - -if (!TPM_IS_CRB (mr->owner)) { -return false; - } - -/* this is a known safe misaligned region, just trace for debug purpose */ -trace_vfio_known_safe_misalignment (memory_region_name (mr), - section->offset_within_address_space, - section->offset_within_region, -qemu_real_host_page_size ()); -return true; +return (section->offset_within_address_space & ~qemu_real_host_page_mask ()) != + (section->offset_within_region & ~qemu_real_host_page_mask ()); } static bool vfio_listener_valid_section(MemoryRegionSection *section,
@@ -379,10 +368,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section,return false; } -if (unlikely ((section->offset_within_address_space & - ~qemu_real_host_page_mask ()) != - (section->offset_within_region & ~qemu_real_host_page_mask ()))) { -if (!vfio_known_safe_misalignment (section)) { +if (unlikely (vfio_section_misaligned (section))) { +/* +* Only warn for VFIO device regions. Other VFIO backends +* (AP, CCW) don't expose memory regions through this path. +*/ +if (vfio_get_vfio_device (memory_region_owner (section->mr))) { error_report("%s received unaligned region %s iova=0x%"PRIx64 " offset_within_region=0x%"PRIx64 " qemu_real_host_page_size=0x%"PRIxPTR,
@@ -390,6 +381,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section, section->offset_within_address_space, section->offset_within_region, qemu_real_host_page_size()); + }else { +trace_vfio_listener_region_misaligned ( +memory_region_name (section->mr), + section->offset_within_address_space, + section->offset_within_region, +qemu_real_host_page_size ()); }return false; }
hw/vfio/trace-events +1 -1
@@ -98,7 +98,7 @@ vfio_spapr_group_attach(int groupfd, int tablefd) "Attached groupfd %d to liobn vfio_listener_region_add_iommu(const char* name, uint64_t start, uint64_t end) "region_add [iommu] %s 0x%"PRIx64" - 0x%"PRIx64 vfio_listener_region_del_iommu(const char *name) "region_del [iommu] %s" vfio_listener_region_add_ram(uint64_t iova_start, uint64_t iova_end, void *vaddr) "region_add [ram] 0x%"PRIx64" - 0x%"PRIx64" [%p]" -vfio_known_safe_misalignment(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR +vfio_listener_region_misaligned(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR vfio_listener_region_add_no_dma_map(const char *name, uint64_t iova, uint64_t size, uint64_t page_size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" is not aligned to 0x%"PRIx64" and cannot be mapped for DMA" vfio_listener_region_skip_dma_map(const char *name, uint64_t iova, uint64_t size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" marked to skip IOMMU mapping" vfio_listener_region_del(uint64_t start, uint64_t end) "region_del 0x%"PRIx64" - 0x%"PRIx64
Signature · 1 line
2.54.0
Hello !
Quoted 14 lines · Cédric Le Goater
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.
Quoted 7 lines
Suggested-by Peter Maydell <[email protected]>Signed-off-by Cédric Le Goater <[email protected]>--- hw/vfio/device.c | 2 +- hw/vfio/listener.c | 33 +++++++++++++++------------------ hw/vfio/trace-events | 2 +- 3 files changed, 17 insertions(+), 20 deletions(-)
Could someone please take a look and Ack this change (or not :) ?
Thanks,
C.
Quoted 93 lines
hw/vfio/device.c +1
-1@@ -439,7 +439,7 @@ bool vfio_device_hiod_create_and_realize(VFIODevice *vbasedev,VFIODevice *vfio_get_vfio_device (Object *obj) { -if (object_dynamic_cast (obj,TYPE_VFIO_PCI )) { +if (object_dynamic_cast (obj,TYPE_VFIO_PCI_DEVICE )) {return &VFIO_PCI_DEVICE (obj)->vbasedev; }else {return NULL ;hw/vfio/listener.c +15
-18@@ -40,7 +40,6 @@#include "migration/misc.h" #include "migration/qemu-file.h" #include "system/tcg.h" -#include "system/tpm.h" #include "vfio-migration-internal.h" #include "vfio-helpers.h" #include "vfio-listener.h" @@ -352,20 +351,10 @@ static void vfio_ram_discard_unregister_listener(VFIOContainer *bcontainer,g_free (vrdl); } -static bool vfio_known_safe_misalignment (MemoryRegionSection *section) +static bool vfio_section_misaligned (MemoryRegionSection *section) { -MemoryRegion *mr = section->mr; - -if (!TPM_IS_CRB (mr->owner)) { -return false; - } - -/* this is a known safe misaligned region, just trace for debug purpose */ -trace_vfio_known_safe_misalignment (memory_region_name (mr), - section->offset_within_address_space, - section->offset_within_region, -qemu_real_host_page_size ()); -return true; +return (section->offset_within_address_space & ~qemu_real_host_page_mask ()) != + (section->offset_within_region & ~qemu_real_host_page_mask ()); } static bool vfio_listener_valid_section(MemoryRegionSection *section,@@ -379,10 +368,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section,return false; } -if (unlikely ((section->offset_within_address_space & - ~qemu_real_host_page_mask ()) != - (section->offset_within_region & ~qemu_real_host_page_mask ()))) { -if (!vfio_known_safe_misalignment (section)) { +if (unlikely (vfio_section_misaligned (section))) { +/* +* Only warn for VFIO device regions. Other VFIO backends +* (AP, CCW) don't expose memory regions through this path. +*/ +if (vfio_get_vfio_device (memory_region_owner (section->mr))) { error_report("%s received unaligned region %s iova=0x%"PRIx64 " offset_within_region=0x%"PRIx64 " qemu_real_host_page_size=0x%"PRIxPTR,@@ -390,6 +381,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section, section->offset_within_address_space, section->offset_within_region, qemu_real_host_page_size()); + }else { +trace_vfio_listener_region_misaligned ( +memory_region_name (section->mr), + section->offset_within_address_space, + section->offset_within_region, +qemu_real_host_page_size ()); }return false; }hw/vfio/trace-events +1
-1@@ -98,7 +98,7 @@ vfio_spapr_group_attach(int groupfd, int tablefd) "Attached groupfd %d to liobn vfio_listener_region_add_iommu(const char* name, uint64_t start, uint64_t end) "region_add [iommu] %s 0x%"PRIx64" - 0x%"PRIx64 vfio_listener_region_del_iommu(const char *name) "region_del [iommu] %s" vfio_listener_region_add_ram(uint64_t iova_start, uint64_t iova_end, void *vaddr) "region_add [ram] 0x%"PRIx64" - 0x%"PRIx64" [%p]" -vfio_known_safe_misalignment(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR +vfio_listener_region_misaligned(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR vfio_listener_region_add_no_dma_map(const char *name, uint64_t iova, uint64_t size, uint64_t page_size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" is not aligned to 0x%"PRIx64" and cannot be mapped for DMA" vfio_listener_region_skip_dma_map(const char *name, uint64_t iova, uint64_t size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" marked to skip IOMMU mapping" vfio_listener_region_del(uint64_t start, uint64_t end) "region_del 0x%"PRIx64" - 0x%"PRIx64
Quoted 10 lines · On Thu, 2 Jul 2026 18:06:40 +0200 Cédric Le Goater <[email protected]> wrote:
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.
Quoted 3 lines
Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.
I thought the whole point was to warn when there are non-vfio devices that we cannot insert into the DMA map. 851d6d1a0ff2 added tpm-crb-cmd to an ignore list because it's not a DMA target. The intention of that ignore list was to continue to evaluate devices that are misaligned and either add them to the ignore list or determine they could be a valid DMA target and correct the alignment.
Can this even legitimately trigger with a vfio-pci device? Thanks,
Alex
Quoted 101 lines
Suggested-by Peter Maydell <[email protected]>Signed-off-by Cédric Le Goater <[email protected]>--- hw/vfio/device.c | 2 +- hw/vfio/listener.c | 33 +++++++++++++++------------------ hw/vfio/trace-events | 2 +- 3 files changed, 17 insertions(+), 20 deletions(-)hw/vfio/device.c +1
-1@@ -439,7 +439,7 @@ bool vfio_device_hiod_create_and_realize(VFIODevice *vbasedev,VFIODevice *vfio_get_vfio_device (Object *obj) { -if (object_dynamic_cast (obj,TYPE_VFIO_PCI )) { +if (object_dynamic_cast (obj,TYPE_VFIO_PCI_DEVICE )) {return &VFIO_PCI_DEVICE (obj)->vbasedev; }else {return NULL ;hw/vfio/listener.c +15
-18@@ -40,7 +40,6 @@#include "migration/misc.h" #include "migration/qemu-file.h" #include "system/tcg.h" -#include "system/tpm.h" #include "vfio-migration-internal.h" #include "vfio-helpers.h" #include "vfio-listener.h" @@ -352,20 +351,10 @@ static void vfio_ram_discard_unregister_listener(VFIOContainer *bcontainer,g_free (vrdl); } -static bool vfio_known_safe_misalignment (MemoryRegionSection *section) +static bool vfio_section_misaligned (MemoryRegionSection *section) { -MemoryRegion *mr = section->mr; - -if (!TPM_IS_CRB (mr->owner)) { -return false; - } - -/* this is a known safe misaligned region, just trace for debug purpose */ -trace_vfio_known_safe_misalignment (memory_region_name (mr), - section->offset_within_address_space, - section->offset_within_region, -qemu_real_host_page_size ()); -return true; +return (section->offset_within_address_space & ~qemu_real_host_page_mask ()) != + (section->offset_within_region & ~qemu_real_host_page_mask ()); } static bool vfio_listener_valid_section(MemoryRegionSection *section,@@ -379,10 +368,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section,return false; } -if (unlikely ((section->offset_within_address_space & - ~qemu_real_host_page_mask ()) != - (section->offset_within_region & ~qemu_real_host_page_mask ()))) { -if (!vfio_known_safe_misalignment (section)) { +if (unlikely (vfio_section_misaligned (section))) { +/* +* Only warn for VFIO device regions. Other VFIO backends +* (AP, CCW) don't expose memory regions through this path. +*/ +if (vfio_get_vfio_device (memory_region_owner (section->mr))) { error_report("%s received unaligned region %s iova=0x%"PRIx64 " offset_within_region=0x%"PRIx64 " qemu_real_host_page_size=0x%"PRIxPTR,@@ -390,6 +381,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section, section->offset_within_address_space, section->offset_within_region, qemu_real_host_page_size()); + }else { +trace_vfio_listener_region_misaligned ( +memory_region_name (section->mr), + section->offset_within_address_space, + section->offset_within_region, +qemu_real_host_page_size ()); }return false; }hw/vfio/trace-events +1
-1@@ -98,7 +98,7 @@ vfio_spapr_group_attach(int groupfd, int tablefd) "Attached groupfd %d to liobn vfio_listener_region_add_iommu(const char* name, uint64_t start, uint64_t end) "region_add [iommu] %s 0x%"PRIx64" - 0x%"PRIx64 vfio_listener_region_del_iommu(const char *name) "region_del [iommu] %s" vfio_listener_region_add_ram(uint64_t iova_start, uint64_t iova_end, void *vaddr) "region_add [ram] 0x%"PRIx64" - 0x%"PRIx64" [%p]" -vfio_known_safe_misalignment(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR +vfio_listener_region_misaligned(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR vfio_listener_region_add_no_dma_map(const char *name, uint64_t iova, uint64_t size, uint64_t page_size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" is not aligned to 0x%"PRIx64" and cannot be mapped for DMA" vfio_listener_region_skip_dma_map(const char *name, uint64_t iova, uint64_t size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" marked to skip IOMMU mapping" vfio_listener_region_del(uint64_t start, uint64_t end) "region_del 0x%"PRIx64" - 0x%"PRIx64
Quoted 17 lines · Alex Williamson
Quoted 10 lines · On Thu, 2 Jul 2026 18:06:40 +0200 Cédric Le Goater <[email protected]> wrote:
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.Quoted 3 lines
Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.
Quoted 6 lines
I thought the whole point was to warn when there are non-vfio devices that we cannot insert into the DMA map.
851d6d1a0ff2added tpm-crb-cmd to an ignore list because it's not a DMA target. The intention of that ignore list was to continue to evaluate devices that are misaligned and either add them to the ignore list or determine they could be a valid DMA target and correct the alignment.
If the goal was to use the VFIO listener to catch all misaligned regions, and potentially fix them, we should reconsider that approach :
https://lore.kernel.org/qemu-devel/[email protected]/
As Peter pointed out, regions that are not page-aligned regions are valid :
and aligning them would require a machine compat for migration. Not addressed in :
https://lore.kernel.org/qemu-devel/[email protected]/
Quoted 1 line
Can this even legitimately trigger with a vfio-pci device?
This would be a real error to report. Right ?
Thanks,
C.
Quoted 106 lines
Thanks,
Alex
Quoted 101 lines
Suggested-by Peter Maydell <[email protected]>Signed-off-by Cédric Le Goater <[email protected]>--- hw/vfio/device.c | 2 +- hw/vfio/listener.c | 33 +++++++++++++++------------------ hw/vfio/trace-events | 2 +- 3 files changed, 17 insertions(+), 20 deletions(-)hw/vfio/device.c +1
-1@@ -439,7 +439,7 @@ bool vfio_device_hiod_create_and_realize(VFIODevice *vbasedev,VFIODevice *vfio_get_vfio_device (Object *obj) { -if (object_dynamic_cast (obj,TYPE_VFIO_PCI )) { +if (object_dynamic_cast (obj,TYPE_VFIO_PCI_DEVICE )) {return &VFIO_PCI_DEVICE (obj)->vbasedev; }else {return NULL ;hw/vfio/listener.c +15
-18@@ -40,7 +40,6 @@#include "migration/misc.h" #include "migration/qemu-file.h" #include "system/tcg.h" -#include "system/tpm.h" #include "vfio-migration-internal.h" #include "vfio-helpers.h" #include "vfio-listener.h" @@ -352,20 +351,10 @@ static void vfio_ram_discard_unregister_listener(VFIOContainer *bcontainer,g_free (vrdl); } -static bool vfio_known_safe_misalignment (MemoryRegionSection *section) +static bool vfio_section_misaligned (MemoryRegionSection *section) { -MemoryRegion *mr = section->mr; - -if (!TPM_IS_CRB (mr->owner)) { -return false; - } - -/* this is a known safe misaligned region, just trace for debug purpose */ -trace_vfio_known_safe_misalignment (memory_region_name (mr), - section->offset_within_address_space, - section->offset_within_region, -qemu_real_host_page_size ()); -return true; +return (section->offset_within_address_space & ~qemu_real_host_page_mask ()) != + (section->offset_within_region & ~qemu_real_host_page_mask ()); } static bool vfio_listener_valid_section(MemoryRegionSection *section,@@ -379,10 +368,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section,return false; } -if (unlikely ((section->offset_within_address_space & - ~qemu_real_host_page_mask ()) != - (section->offset_within_region & ~qemu_real_host_page_mask ()))) { -if (!vfio_known_safe_misalignment (section)) { +if (unlikely (vfio_section_misaligned (section))) { +/* +* Only warn for VFIO device regions. Other VFIO backends +* (AP, CCW) don't expose memory regions through this path. +*/ +if (vfio_get_vfio_device (memory_region_owner (section->mr))) { error_report("%s received unaligned region %s iova=0x%"PRIx64 " offset_within_region=0x%"PRIx64 " qemu_real_host_page_size=0x%"PRIxPTR,@@ -390,6 +381,12 @@ static bool vfio_listener_valid_section(MemoryRegionSection *section, section->offset_within_address_space, section->offset_within_region, qemu_real_host_page_size()); + }else { +trace_vfio_listener_region_misaligned ( +memory_region_name (section->mr), + section->offset_within_address_space, + section->offset_within_region, +qemu_real_host_page_size ()); }return false; }hw/vfio/trace-events +1
-1@@ -98,7 +98,7 @@ vfio_spapr_group_attach(int groupfd, int tablefd) "Attached groupfd %d to liobn vfio_listener_region_add_iommu(const char* name, uint64_t start, uint64_t end) "region_add [iommu] %s 0x%"PRIx64" - 0x%"PRIx64 vfio_listener_region_del_iommu(const char *name) "region_del [iommu] %s" vfio_listener_region_add_ram(uint64_t iova_start, uint64_t iova_end, void *vaddr) "region_add [ram] 0x%"PRIx64" - 0x%"PRIx64" [%p]" -vfio_known_safe_misalignment(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR +vfio_listener_region_misaligned(const char *name, uint64_t iova, uint64_t offset_within_region, uintptr_t page_size) "Region \"%s\" iova=0x%"PRIx64" offset_within_region=0x%"PRIx64" qemu_real_host_page_size=0x%"PRIxPTR vfio_listener_region_add_no_dma_map(const char *name, uint64_t iova, uint64_t size, uint64_t page_size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" is not aligned to 0x%"PRIx64" and cannot be mapped for DMA" vfio_listener_region_skip_dma_map(const char *name, uint64_t iova, uint64_t size) "Region \"%s\" 0x%"PRIx64" size=0x%"PRIx64" marked to skip IOMMU mapping" vfio_listener_region_del(uint64_t start, uint64_t end) "region_del 0x%"PRIx64" - 0x%"PRIx64
Quoted 29 lines · Cédric Le Goater
Quoted 17 lines · Alex Williamson
Quoted 10 lines · On Thu, 2 Jul 2026 18:06:40 +0200 Cédric Le Goater <[email protected]> wrote:
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.Quoted 3 lines
Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.Quoted 6 lines
I thought the whole point was to warn when there are non-vfio devices that we cannot insert into the DMA map.
851d6d1a0ff2added tpm-crb-cmd to an ignore list because it's not a DMA target. The intention of that ignore list was to continue to evaluate devices that are misaligned and either add them to the ignore list or determine they could be a valid DMA target and correct the alignment.If the goal was to use the VFIO listener to catch all misaligned regions, and potentially fix them, we should reconsider that approach :
Quoted 7 lines
https://lore.kernel.org/qemu-devel/[email protected]/As Peter pointed out, regions that are not page-aligned regions are valid :
and aligning them would require a machine compat for migration.
It seems very unlikely that anything not page-aligned is going to be a sensible target for DMA ops from a device: the main reason to have a non-page-aligned or non-page-sized region is that it's the way we've chosen to implement a device, or it's some "RAM but not general-purpose RAM" thing. If there's a path to warn/error if it's actually used for a DMA op that would be worthwhile, but as it stands I think all this warning will ever cause us to do is add more entries to its whitelist every time it complains.
thanks
-- PMM
Quoted 36 lines · On Mon, 6 Jul 2026 16:27:46 +0100 Peter Maydell <[email protected]>…
Quoted 35 lines · Cédric Le Goater
Quoted 17 lines · Alex Williamson
Quoted 10 lines · On Thu, 2 Jul 2026 18:06:40 +0200 Cédric Le Goater <[email protected]> wrote:
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.Quoted 3 lines
Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.Quoted 6 lines
I thought the whole point was to warn when there are non-vfio devices that we cannot insert into the DMA map.
851d6d1a0ff2added tpm-crb-cmd to an ignore list because it's not a DMA target. The intention of that ignore list was to continue to evaluate devices that are misaligned and either add them to the ignore list or determine they could be a valid DMA target and correct the alignment.If the goal was to use the VFIO listener to catch all misaligned regions, and potentially fix them, we should reconsider that approach :
https://lore.kernel.org/qemu-devel/[email protected]/As Peter pointed out, regions that are not page-aligned regions are valid :
Quoted 8 lines
Quoted 1 line
and aligning them would require a machine compat for migration.
It seems very unlikely that anything not page-aligned is going to be a sensible target for DMA ops from a device: the main reason to have a non-page-aligned or non-page-sized region is that it's the way we've chosen to implement a device, or it's some "RAM but not general-purpose RAM" thing. If there's a path to warn/error if it's actually used for a DMA op that would be worthwhile,
But for a vfio-pci device there is no DMA op, it either gets mapped into the IOMMU for direct access to the device or not. This is the point at which that decision is made.
Quoted 2 lines
but as it stands I think all this warning will ever cause us to do is add more entries to its whitelist every time it complains.
IMO, that's ok. That's the purpose. We're up to one whole class of device so far, we don't care about mapping failures to TPM devices. If some day this does fire on something we care about, especially a regression, it'll earn its keep. It doesn't seem like an excessive burden yet. Thanks,
Alex
Quoted 10 lines · Alex Williamson
Quoted 6 lines · On Mon, 6 Jul 2026 16:27:46 +0100 Peter Maydell <[email protected]>…
It seems very unlikely that anything not page-aligned is going to be a sensible target for DMA ops from a device: the main reason to have a non-page-aligned or non-page-sized region is that it's the way we've chosen to implement a device, or it's some "RAM but not general-purpose RAM" thing. If there's a path to warn/error if it's actually used for a DMA op that would be worthwhile,
Quoted 3 lines
But for a vfio-pci device there is no DMA op, it either gets mapped into the IOMMU for direct access to the device or not. This is the point at which that decision is made.
Why is small RAM different from MMIO register regions here? Those also you're not going to be able to have a device do direct memory access to.
Quoted 8 lines
Quoted 2 lines
but as it stands I think all this warning will ever cause us to do is add more entries to its whitelist every time it complains.
IMO, that's ok. That's the purpose. We're up to one whole class of device so far, we don't care about mapping failures to TPM devices. If some day this does fire on something we care about, especially a regression, it'll earn its keep. It doesn't seem like an excessive burden yet. Thanks,
The reason we're having this conversation is because it just fired again, and the warning gives the impression that the right thing to do is to somehow change the misaligned memory region, not to add something to the whitelist. That's already wasted several people's time.
You could add a comment saying
/*
* If this warning ever fires, then it will almost certainly
* be because of a legitimately non-aligned RAM-backed
* MemoryRegion that needs adding to the whitelist. */
But I think that it would be much better just to dump the warning, so we don't have to keep doing this.
thanks
-- PMM
Quoted 11 lines · On Mon, 6 Jul 2026 17:12:35 +0100 Peter Maydell <[email protected]>…
Quoted 10 lines · Alex Williamson
Quoted 6 lines · On Mon, 6 Jul 2026 16:27:46 +0100 Peter Maydell <[email protected]>…
It seems very unlikely that anything not page-aligned is going to be a sensible target for DMA ops from a device: the main reason to have a non-page-aligned or non-page-sized region is that it's the way we've chosen to implement a device, or it's some "RAM but not general-purpose RAM" thing. If there's a path to warn/error if it's actually used for a DMA op that would be worthwhile,
Quoted 7 lines
Quoted 3 lines
But for a vfio-pci device there is no DMA op, it either gets mapped into the IOMMU for direct access to the device or not. This is the point at which that decision is made.
Why is small RAM different from MMIO register regions here? Those also you're not going to be able to have a device do direct memory access to.
There's a precedent for bare metal, peer-to-peer DMA is less reliable and drivers don't typically have blind faith in it working. Peer-to-peer DMA also has specific use cases, drivers don't arbitrarily pick a peer device to DMA to, they have specific use cases and target regions are inherently sized and aligned appropriately.
Here we have RAM and we're trying to determine whether it's RAM that could be allocated as a DMA buffer or expect a DMA access from a device.
Commit 851d6d1a0ff2 spelled out that we could have declared these sorts of regions as RAM device, solving the problem with a simple test, but that wasn't well received (for reasons I don't remember), so we were forced to create this mechanism and maintenance issue.
Quoted 9 lines
Quoted 8 lines
Quoted 2 lines
but as it stands I think all this warning will ever cause us to do is add more entries to its whitelist every time it complains.
IMO, that's ok. That's the purpose. We're up to one whole class of device so far, we don't care about mapping failures to TPM devices. If some day this does fire on something we care about, especially a regression, it'll earn its keep. It doesn't seem like an excessive burden yet. Thanks,
Quoted 5 lines
The reason we're having this conversation is because it just fired again, and the warning gives the impression that the right thing to do is to somehow change the misaligned memory region, not to add something to the whitelist. That's already wasted several people's time.
Since being introduced 4 years ago, we've iterated on one class of device. It doesn't feel like there's a persistent time sink here.
Quoted 7 lines
You could add a comment saying
/*
* If this warning ever fires, then it will almost certainly
* be because of a legitimately non-aligned RAM-backed
* MemoryRegion that needs adding to the whitelist. */
Personally I'd start with the error log itself, providing a hint that the region is presented as RAM but isn't sized/aligned for DMA from a vfio device, with a hint to mark it as safe if it is legitimately not a valid DMA target. Thanks,
Alex
Quoted 23 lines · Alex Williamson
Quoted 11 lines · On Mon, 6 Jul 2026 17:12:35 +0100 Peter Maydell <[email protected]>…
Quoted 10 lines · Alex Williamson
Quoted 6 lines · On Mon, 6 Jul 2026 16:27:46 +0100 Peter Maydell <[email protected]>…
It seems very unlikely that anything not page-aligned is going to be a sensible target for DMA ops from a device: the main reason to have a non-page-aligned or non-page-sized region is that it's the way we've chosen to implement a device, or it's some "RAM but not general-purpose RAM" thing. If there's a path to warn/error if it's actually used for a DMA op that would be worthwhile,
Quoted 7 lines
Quoted 3 lines
But for a vfio-pci device there is no DMA op, it either gets mapped into the IOMMU for direct access to the device or not. This is the point at which that decision is made.
Why is small RAM different from MMIO register regions here? Those also you're not going to be able to have a device do direct memory access to.
Quoted 5 lines
There's a precedent for bare metal, peer-to-peer DMA is less reliable and drivers don't typically have blind faith in it working. Peer-to-peer DMA also has specific use cases, drivers don't arbitrarily pick a peer device to DMA to, they have specific use cases and target regions are inherently sized and aligned appropriately.
Sure, but if a guest does try it then it's going to fall over. This is exactly the same as if it tries to do DMA to a small MemoryRegion that we happen to have implemented as backed by RAM rather than MMIO functions. In both cases the QEMU implementation of the device is fine, and the guest has done something silly.
Quoted 3 lines
Here we have RAM and we're trying to determine whether it's RAM that could be allocated as a DMA buffer or expect a DMA access from a device.
I think that there are no examples of memory-backed MemoryRegions that aren't page-aligned that are going to expect a DMA access. So a reasonable proxy test for "is this RAM that could expect a DMA access" is "is it page aligned and at least a page in size?". All blocks of real for-the-guest RAM will satisfy that.
Quoted 4 lines
Commit
851d6d1a0ff2spelled out that we could have declared these sorts of regions as RAM device, solving the problem with a simple test, but that wasn't well received (for reasons I don't remember), so we were forced to create this mechanism and maintenance issue.
Isn't a RAM device for "this MemoryRegion is backed by memory from some actual host hardware device"? In this case it is not, so not a good fit. But either way, if the RAM-device MR is not page aligned then the guest can't DMA to it either. So "is this a RAM device or not" doesn't seem like the test you want for whether you want to flag a problem.
thanks
-- PMM
Quoted 25 lines · On Mon, 6 Jul 2026 17:12:07 +0200 Cédric Le Goater <[email protected]> wrote:
Quoted 17 lines · Alex Williamson
Quoted 10 lines · On Thu, 2 Jul 2026 18:06:40 +0200 Cédric Le Goater <[email protected]> wrote:
Any device on the board can have non-page-aligned memory regions. The VFIO listener should not warn about misalignment for regions that are not VFIO-related.
Replace the
vfio_known_safe_misalignment()whitelist with avfio_get_vfio_device()ownership check: only warn when a VFIO device's own region is misaligned. For all other regions, emit a trace event and silently skip. Extract the misalignment test into avfio_section_misaligned()helper for readability.Quoted 3 lines
Note:
vfio_get_vfio_device()currently only covers VFIO PCI devices, including vfio-user-pci. Other VFIO backends (AP, CCW) don't expose memory regions through this path so this is sufficient for now.Quoted 6 lines
I thought the whole point was to warn when there are non-vfio devices that we cannot insert into the DMA map.
851d6d1a0ff2added tpm-crb-cmd to an ignore list because it's not a DMA target. The intention of that ignore list was to continue to evaluate devices that are misaligned and either add them to the ignore list or determine they could be a valid DMA target and correct the alignment.
Quoted 4 lines
If the goal was to use the VFIO listener to catch all misaligned regions, and potentially fix them, we should reconsider that approach :
https://lore.kernel.org/qemu-devel/[email protected]/
Arguably ok, likely no tpm device is a useful DMA target.
Quoted 3 lines
As Peter pointed out, regions that are not page-aligned regions are valid :
Yes, they are valid, no argument. We're trying to DMA map things. We can't DMA map things that aren't page-aligned and sized. If it's not likely a DMA target, well, that's the purpose of the ignore list. If it might be a DMA target, this is intended to log why the VM may not work correctly vs bare metal.
Quoted 4 lines
and aligning them would require a machine compat for migration. Not addressed in :
https://lore.kernel.org/qemu-devel/[email protected]/
Which I think highlights why we might rather see warnings than bury the issue in traces.
Quoted 3 lines
Quoted 1 line
Can this even legitimately trigger with a vfio-pci device?
This would be a real error to report. Right ?
I suspect it's not possible, so as proposed here we're turning a blind eye to the issue we intended to monitor and validate, and we replace it with likely dead code. Thanks,
Alex