Re: [PATCH v2 1/1] vhost-vdpa: mute unaligned memory error report

[Laurent Vivier <lvivier@redhat.com>]

On 7/4/23 09:25, David Hildenbrand wrote:
> On 04.07.23 09:19, Laurent Vivier wrote:
>> With TPM CRM device, vhost-vdpa reports an error when it tries
>> to register a listener for a non aligned memory region:
>>
>>    qemu-system-x86_64: vhost_vdpa_listener_region_add received unaligned region
>>    qemu-system-x86_64: vhost_vdpa_listener_region_del received unaligned region
>>
>> This error can be confusing for the user whereas we only need to skip
>> the region (as it's already done after the error_report())
>>
>> Rather than introducing a special case for TPM CRB memory section
>> to not display the message in this case, simply replace the
>> error_report() by a trace function (with more information, like the
>> memory region name).
>>
>> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
>> ---
>>   hw/virtio/trace-events | 2 ++
>>   hw/virtio/vhost-vdpa.c | 8 ++++++--
>>   2 files changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/hw/virtio/trace-events b/hw/virtio/trace-events
>> index 8f8d05cf9b01..9b0d643b9475 100644
>> --- a/hw/virtio/trace-events
>> +++ b/hw/virtio/trace-events
>> @@ -34,7 +34,9 @@ vhost_vdpa_dma_map(void *vdpa, int fd, uint32_t msg_type, uint32_t 
>> asid, uint64_
>>   vhost_vdpa_dma_unmap(void *vdpa, int fd, uint32_t msg_type, uint32_t asid, uint64_t 
>> iova, uint64_t size, uint8_t type) "vdpa:%p fd: %d msg_type: %"PRIu32" asid: %"PRIu32" 
>> iova: 0x%"PRIx64" size: 0x%"PRIx64" type: %"PRIu8
>>   vhost_vdpa_listener_begin_batch(void *v, int fd, uint32_t msg_type, uint8_t type)  
>> "vdpa:%p fd: %d msg_type: %"PRIu32" type: %"PRIu8
>>   vhost_vdpa_listener_commit(void *v, int fd, uint32_t msg_type, uint8_t type)  "vdpa:%p 
>> fd: %d msg_type: %"PRIu32" type: %"PRIu8
>> +vhost_vdpa_listener_region_add_unaligned(void *v, const char *name, uint64_t offset_as, 
>> uint64_t offset_page) "vdpa: %p region %s offset_within_address_space %"PRIu64" 
>> offset_within_region %"PRIu64
>>   vhost_vdpa_listener_region_add(void *vdpa, uint64_t iova, uint64_t llend, void *vaddr, 
>> bool readonly) "vdpa: %p iova 0x%"PRIx64" llend 0x%"PRIx64" vaddr: %p read-only: %d"
>> +vhost_vdpa_listener_region_del_unaligned(void *v, const char *name, uint64_t offset_as, 
>> uint64_t offset_page) "vdpa: %p region %s offset_within_address_space %"PRIu64" 
>> offset_within_region %"PRIu64
>>   vhost_vdpa_listener_region_del(void *vdpa, uint64_t iova, uint64_t llend) "vdpa: %p 
>> iova 0x%"PRIx64" llend 0x%"PRIx64
>>   vhost_vdpa_add_status(void *dev, uint8_t status) "dev: %p status: 0x%"PRIx8
>>   vhost_vdpa_init(void *dev, void *vdpa) "dev: %p vdpa: %p"
>> diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
>> index 3c575a9a6e9e..24d32f0d3728 100644
>> --- a/hw/virtio/vhost-vdpa.c
>> +++ b/hw/virtio/vhost-vdpa.c
>> @@ -323,7 +323,9 @@ static void vhost_vdpa_listener_region_add(MemoryListener *listener,
>>       if (unlikely((section->offset_within_address_space & ~TARGET_PAGE_MASK) !=
>>                    (section->offset_within_region & ~TARGET_PAGE_MASK))) {
>> -        error_report("%s received unaligned region", __func__);
>> +        trace_vhost_vdpa_listener_region_add_unaligned(v, section->mr->name,
>> +                       section->offset_within_address_space & ~TARGET_PAGE_MASK,
>> +                       section->offset_within_region & ~TARGET_PAGE_MASK);
>>           return;
>>       }
>> @@ -405,7 +407,9 @@ static void vhost_vdpa_listener_region_del(MemoryListener *listener,
>>       if (unlikely((section->offset_within_address_space & ~TARGET_PAGE_MASK) !=
>>                    (section->offset_within_region & ~TARGET_PAGE_MASK))) {
>> -        error_report("%s received unaligned region", __func__);
>> +        trace_vhost_vdpa_listener_region_del_unaligned(v, section->mr->name,
>> +                       section->offset_within_address_space & ~TARGET_PAGE_MASK,
>> +                       section->offset_within_region & ~TARGET_PAGE_MASK);
>>           return;
>>       }
> 
> Reviewed-by: David Hildenbrand <david@redhat.com>
> 
> 
> Do we also want to touch the vfio side in vfio_listener_valid_section(), or why is that 
> one unaffected?
> 

I don't know if we can apply the same solution for VFIO.
I don't know if the error message is relevant or if we can keep only the trace and remove 
the error_report() for all the cases (in this case vfio_known_safe_misalignment() becomes 
useless).

Thanks,
Laurent