Re: [PATCH 3/3] xen/virtio: enable grant based virtio on x86

[Juergen Gross <jgross@suse.com>]

[-- Attachment #1.1.1: Type: text/plain, Size: 3315 bytes --]

On 06.10.22 18:04, Oleksandr Tyshchenko wrote:
> 
> On 06.10.22 10:15, Juergen Gross wrote:
> 
> 
> Hello Juergen
> 
>> Use an x86-specific virtio_check_mem_acc_cb() for Xen in order to setup
>> the correct DMA ops.
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
>> ---
>>    arch/x86/xen/enlighten_hvm.c |  2 +-
>>    arch/x86/xen/enlighten_pv.c  |  2 +-
>>    drivers/xen/grant-dma-ops.c  | 10 ++++++++++
>>    include/xen/xen-ops.h        |  1 +
>>    4 files changed, 13 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/x86/xen/enlighten_hvm.c b/arch/x86/xen/enlighten_hvm.c
>> index 1c1ac418484b..c1cd28e915a3 100644
>> --- a/arch/x86/xen/enlighten_hvm.c
>> +++ b/arch/x86/xen/enlighten_hvm.c
>> @@ -212,7 +212,7 @@ static void __init xen_hvm_guest_init(void)
>>    		return;
>>    
>>    	if (IS_ENABLED(CONFIG_XEN_VIRTIO_FORCE_GRANT))
>> -		virtio_set_mem_acc_cb(virtio_require_restricted_mem_acc);
>> +		virtio_set_mem_acc_cb(xen_virtio_restricted_mem_acc);
>>    
>>    	init_hvm_pv_info();
>>    
>> diff --git a/arch/x86/xen/enlighten_pv.c b/arch/x86/xen/enlighten_pv.c
>> index 9b1a58dda935..45b24c1b646a 100644
>> --- a/arch/x86/xen/enlighten_pv.c
>> +++ b/arch/x86/xen/enlighten_pv.c
>> @@ -112,7 +112,7 @@ static void __init xen_pv_init_platform(void)
>>    {
>>    	/* PV guests can't operate virtio devices without grants. */
>>    	if (IS_ENABLED(CONFIG_XEN_VIRTIO))
>> -		virtio_set_mem_acc_cb(virtio_require_restricted_mem_acc);
>> +		virtio_set_mem_acc_cb(xen_virtio_restricted_mem_acc);
>>    
>>    	populate_extra_pte(fix_to_virt(FIX_PARAVIRT_BOOTMAP));
>>    
>> diff --git a/drivers/xen/grant-dma-ops.c b/drivers/xen/grant-dma-ops.c
>> index a00112235877..60a7acc334ed 100644
>> --- a/drivers/xen/grant-dma-ops.c
>> +++ b/drivers/xen/grant-dma-ops.c
>> @@ -372,6 +372,16 @@ void xen_grant_setup_dma_ops(struct device *dev)
>>    	dev_err(dev, "Cannot set up Xen grant DMA ops, retain platform DMA ops\n");
>>    }
>>    
>> +bool xen_virtio_restricted_mem_acc(struct virtio_device *dev)
>> +{
>> +	bool ret = xen_virtio_mem_acc(dev);
> 
> 
> The grant usage is mandatory for PV guests, right?
> 
> Then xen_virtio_mem_acc() should always return true for PV guests (I
> mean even if CONFIG_XEN_VIRTIO_FORCE_GRANT is not set).

Yes.

> 
> 
> 
>> +
>> +	if (ret)
>> +		xen_grant_setup_dma_ops(dev->dev.parent);
>> +
>> +	return ret;
>> +}
>> +
>>    MODULE_DESCRIPTION("Xen grant DMA-mapping layer");
>>    MODULE_AUTHOR("Juergen Gross <jgross@suse.com>");
>>    MODULE_LICENSE("GPL");
>> diff --git a/include/xen/xen-ops.h b/include/xen/xen-ops.h
>> index dae0f350c678..3dd5aa936f1d 100644
>> --- a/include/xen/xen-ops.h
>> +++ b/include/xen/xen-ops.h
>> @@ -219,6 +219,7 @@ static inline void xen_preemptible_hcall_end(void) { }
>>    void xen_grant_setup_dma_ops(struct device *dev);
>>    bool xen_is_grant_dma_device(struct device *dev);
>>    bool xen_virtio_mem_acc(struct virtio_device *dev);
>> +bool xen_virtio_restricted_mem_acc(struct virtio_device *dev);
>>    #else
>>    static inline void xen_grant_setup_dma_ops(struct device *dev)
>>    {
> 
> 
> And probably static inline stub always returning false if
> CONFIG_XEN_GRANT_DMA_OPS is not set.

Indeed.


Juergen


[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 3149 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 495 bytes --]