Re: [PATCH v3 05/11] vduse: Support automatic irq callback affinity

Re: [PATCH v3 05/11] vduse: Support automatic irq callback affinity

[kernel test robot]

Hi Xie,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on tip/irq/core]
[also build test WARNING on linus/master next-20230228]
[cannot apply to mst-vhost/linux-next v6.2]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Xie-Yongji/lib-group_cpus-Export-group_cpus_evenly/20230228-174438
patch link:    https://lore.kernel.org/r/20230228094110.37-6-xieyongji%40bytedance.com
patch subject: [PATCH v3 05/11] vduse: Support automatic irq callback affinity
config: m68k-allyesconfig (https://download.01.org/0day-ci/archive/20230228/202302281954.jRA7Qzq4-lkp@intel.com/config)
compiler: m68k-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/intel-lab-lkp/linux/commit/6c15cc28cb814c0e6cb80955bc59517e80c15ae2
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Xie-Yongji/lib-group_cpus-Export-group_cpus_evenly/20230228-174438
        git checkout 6c15cc28cb814c0e6cb80955bc59517e80c15ae2
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=m68k olddefconfig
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=m68k SHELL=/bin/bash drivers/vdpa/vdpa_user/

If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
| Link: https://lore.kernel.org/oe-kbuild-all/202302281954.jRA7Qzq4-lkp@intel.com/

All warnings (new ones prefixed by >>):

>> drivers/vdpa/vdpa_user/vduse_dev.c:725:1: warning: no previous prototype for 'create_affinity_masks' [-Wmissing-prototypes]
     725 | create_affinity_masks(unsigned int nvecs, struct irq_affinity *affd)
         | ^~~~~~~~~~~~~~~~~~~~~


vim +/create_affinity_masks +725 drivers/vdpa/vdpa_user/vduse_dev.c

   723	
   724	struct cpumask *
 > 725	create_affinity_masks(unsigned int nvecs, struct irq_affinity *affd)
   726	{
   727		unsigned int affvecs = 0, curvec, usedvecs, i;
   728		struct cpumask *masks = NULL;
   729	
   730		if (nvecs > affd->pre_vectors + affd->post_vectors)
   731			affvecs = nvecs - affd->pre_vectors - affd->post_vectors;
   732	
   733		if (!affd->calc_sets)
   734			affd->calc_sets = default_calc_sets;
   735	
   736		affd->calc_sets(affd, affvecs);
   737	
   738		if (!affvecs)
   739			return NULL;
   740	
   741		masks = kcalloc(nvecs, sizeof(*masks), GFP_KERNEL);
   742		if (!masks)
   743			return NULL;
   744	
   745		/* Fill out vectors at the beginning that don't need affinity */
   746		for (curvec = 0; curvec < affd->pre_vectors; curvec++)
   747			cpumask_setall(&masks[curvec]);
   748	
   749		for (i = 0, usedvecs = 0; i < affd->nr_sets; i++) {
   750			unsigned int this_vecs = affd->set_size[i];
   751			int j;
   752			struct cpumask *result = group_cpus_evenly(this_vecs);
   753	
   754			if (!result) {
   755				kfree(masks);
   756				return NULL;
   757			}
   758	
   759			for (j = 0; j < this_vecs; j++)
   760				cpumask_copy(&masks[curvec + j], &result[j]);
   761			kfree(result);
   762	
   763			curvec += this_vecs;
   764			usedvecs += this_vecs;
   765		}
   766	
   767		/* Fill out vectors at the end that don't need affinity */
   768		if (usedvecs >= affvecs)
   769			curvec = affd->pre_vectors + affvecs;
   770		else
   771			curvec = affd->pre_vectors + usedvecs;
   772		for (; curvec < nvecs; curvec++)
   773			cpumask_setall(&masks[curvec]);
   774	
   775		return masks;
   776	}
   777	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 05/11] vduse: Support automatic irq callback affinity

[kernel test robot]

Hi Xie,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on tip/irq/core]
[also build test WARNING on linus/master next-20230228]
[cannot apply to mst-vhost/linux-next hch-configfs/for-next v6.2]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Xie-Yongji/lib-group_cpus-Export-group_cpus_evenly/20230228-174438
patch link:    https://lore.kernel.org/r/20230228094110.37-6-xieyongji%40bytedance.com
patch subject: [PATCH v3 05/11] vduse: Support automatic irq callback affinity
config: x86_64-randconfig-s021 (https://download.01.org/0day-ci/archive/20230301/202303010802.fyGx4T0d-lkp@intel.com/config)
compiler: gcc-11 (Debian 11.3.0-8) 11.3.0
reproduce:
        # apt-get install sparse
        # sparse version: v0.6.4-39-gce1a6720-dirty
        # https://github.com/intel-lab-lkp/linux/commit/6c15cc28cb814c0e6cb80955bc59517e80c15ae2
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Xie-Yongji/lib-group_cpus-Export-group_cpus_evenly/20230228-174438
        git checkout 6c15cc28cb814c0e6cb80955bc59517e80c15ae2
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=x86_64 olddefconfig
        make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=x86_64 SHELL=/bin/bash drivers/vdpa/vdpa_user/

If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
| Link: https://lore.kernel.org/oe-kbuild-all/202303010802.fyGx4T0d-lkp@intel.com/

sparse warnings: (new ones prefixed by >>)
>> drivers/vdpa/vdpa_user/vduse_dev.c:724:16: sparse: sparse: symbol 'create_affinity_masks' was not declared. Should it be static?

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 05/11] vduse: Support automatic irq callback affinity

[Jason Wang]

在 2023/2/28 17:41, Xie Yongji 写道:
> This brings current interrupt affinity spreading mechanism
> to vduse device. We will make use of group_cpus_evenly()
> to create an irq callback affinity mask for each virtqueue of
> vduse device. Then we will spread IRQs between CPUs in the affinity
> mask, in a round-robin manner, to run the irq callback.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> ---
>   drivers/vdpa/vdpa_user/vduse_dev.c | 130 +++++++++++++++++++++++++++--
>   1 file changed, 123 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/vdpa/vdpa_user/vduse_dev.c b/drivers/vdpa/vdpa_user/vduse_dev.c
> index 98359d87a06f..bde28a8692d5 100644
> --- a/drivers/vdpa/vdpa_user/vduse_dev.c
> +++ b/drivers/vdpa/vdpa_user/vduse_dev.c
> @@ -23,6 +23,8 @@
>   #include <linux/nospec.h>
>   #include <linux/vmalloc.h>
>   #include <linux/sched/mm.h>
> +#include <linux/interrupt.h>
> +#include <linux/group_cpus.h>
>   #include <uapi/linux/vduse.h>
>   #include <uapi/linux/vdpa.h>
>   #include <uapi/linux/virtio_config.h>
> @@ -41,6 +43,8 @@
>   #define VDUSE_IOVA_SIZE (128 * 1024 * 1024)
>   #define VDUSE_MSG_DEFAULT_TIMEOUT 30
>   
> +#define IRQ_UNBOUND -1
> +
>   struct vduse_virtqueue {
>   	u16 index;
>   	u16 num_max;
> @@ -57,6 +61,8 @@ struct vduse_virtqueue {
>   	struct vdpa_callback cb;
>   	struct work_struct inject;
>   	struct work_struct kick;
> +	int irq_effective_cpu;
> +	struct cpumask irq_affinity;
>   };
>   
>   struct vduse_dev;
> @@ -128,6 +134,7 @@ static struct class *vduse_class;
>   static struct cdev vduse_ctrl_cdev;
>   static struct cdev vduse_cdev;
>   static struct workqueue_struct *vduse_irq_wq;
> +static struct workqueue_struct *vduse_irq_bound_wq;
>   
>   static u32 allowed_device_id[] = {
>   	VIRTIO_ID_BLOCK,
> @@ -708,6 +715,82 @@ static u32 vduse_vdpa_get_generation(struct vdpa_device *vdpa)
>   	return dev->generation;
>   }
>   
> +static void default_calc_sets(struct irq_affinity *affd, unsigned int affvecs)
> +{
> +	affd->nr_sets = 1;
> +	affd->set_size[0] = affvecs;
> +}
> +
> +struct cpumask *
> +create_affinity_masks(unsigned int nvecs, struct irq_affinity *affd)
> +{
> +	unsigned int affvecs = 0, curvec, usedvecs, i;
> +	struct cpumask *masks = NULL;
> +
> +	if (nvecs > affd->pre_vectors + affd->post_vectors)
> +		affvecs = nvecs - affd->pre_vectors - affd->post_vectors;
> +
> +	if (!affd->calc_sets)
> +		affd->calc_sets = default_calc_sets;
> +
> +	affd->calc_sets(affd, affvecs);
> +
> +	if (!affvecs)
> +		return NULL;
> +
> +	masks = kcalloc(nvecs, sizeof(*masks), GFP_KERNEL);
> +	if (!masks)
> +		return NULL;
> +
> +	/* Fill out vectors at the beginning that don't need affinity */
> +	for (curvec = 0; curvec < affd->pre_vectors; curvec++)
> +		cpumask_setall(&masks[curvec]);
> +
> +	for (i = 0, usedvecs = 0; i < affd->nr_sets; i++) {
> +		unsigned int this_vecs = affd->set_size[i];
> +		int j;
> +		struct cpumask *result = group_cpus_evenly(this_vecs);
> +
> +		if (!result) {
> +			kfree(masks);
> +			return NULL;
> +		}
> +
> +		for (j = 0; j < this_vecs; j++)
> +			cpumask_copy(&masks[curvec + j], &result[j]);
> +		kfree(result);
> +
> +		curvec += this_vecs;
> +		usedvecs += this_vecs;
> +	}
> +
> +	/* Fill out vectors at the end that don't need affinity */
> +	if (usedvecs >= affvecs)
> +		curvec = affd->pre_vectors + affvecs;
> +	else
> +		curvec = affd->pre_vectors + usedvecs;
> +	for (; curvec < nvecs; curvec++)
> +		cpumask_setall(&masks[curvec]);
> +
> +	return masks;
> +}
> +
> +static void vduse_vdpa_set_irq_affinity(struct vdpa_device *vdpa,
> +					struct irq_affinity *desc)
> +{
> +	struct vduse_dev *dev = vdpa_to_vduse(vdpa);
> +	struct cpumask *masks;
> +	int i;
> +
> +	masks = create_affinity_masks(dev->vq_num, desc);
> +	if (!masks)
> +		return;
> +
> +	for (i = 0; i < dev->vq_num; i++)
> +		cpumask_copy(&dev->vqs[i]->irq_affinity, &masks[i]);
> +	kfree(masks);
> +}
> +
>   static int vduse_vdpa_set_map(struct vdpa_device *vdpa,
>   				unsigned int asid,
>   				struct vhost_iotlb *iotlb)
> @@ -758,6 +841,7 @@ static const struct vdpa_config_ops vduse_vdpa_config_ops = {
>   	.get_config		= vduse_vdpa_get_config,
>   	.set_config		= vduse_vdpa_set_config,
>   	.get_generation		= vduse_vdpa_get_generation,
> +	.set_irq_affinity	= vduse_vdpa_set_irq_affinity,
>   	.reset			= vduse_vdpa_reset,
>   	.set_map		= vduse_vdpa_set_map,
>   	.free			= vduse_vdpa_free,
> @@ -917,7 +1001,8 @@ static void vduse_vq_irq_inject(struct work_struct *work)
>   }
>   
>   static int vduse_dev_queue_irq_work(struct vduse_dev *dev,
> -				    struct work_struct *irq_work)
> +				    struct work_struct *irq_work,
> +				    int irq_effective_cpu)
>   {
>   	int ret = -EINVAL;
>   
> @@ -926,7 +1011,11 @@ static int vduse_dev_queue_irq_work(struct vduse_dev *dev,
>   		goto unlock;
>   
>   	ret = 0;
> -	queue_work(vduse_irq_wq, irq_work);
> +	if (irq_effective_cpu == IRQ_UNBOUND)
> +		queue_work(vduse_irq_wq, irq_work);
> +	else
> +		queue_work_on(irq_effective_cpu,
> +			      vduse_irq_bound_wq, irq_work);
>   unlock:
>   	up_read(&dev->rwsem);
>   
> @@ -1029,6 +1118,22 @@ static int vduse_dev_reg_umem(struct vduse_dev *dev,
>   	return ret;
>   }
>   
> +static void vduse_vq_update_effective_cpu(struct vduse_virtqueue *vq)
> +{
> +	int curr_cpu = vq->irq_effective_cpu;
> +
> +	while (true) {
> +		curr_cpu = cpumask_next(curr_cpu, &vq->irq_affinity);
> +		if (cpu_online(curr_cpu))
> +			break;
> +
> +		if (curr_cpu >= nr_cpu_ids)
> +			curr_cpu = -1;


IRQ_UNBOUND?


> +	}
> +
> +	vq->irq_effective_cpu = curr_cpu;
> +}
> +
>   static long vduse_dev_ioctl(struct file *file, unsigned int cmd,
>   			    unsigned long arg)
>   {
> @@ -1111,7 +1216,7 @@ static long vduse_dev_ioctl(struct file *file, unsigned int cmd,
>   		break;
>   	}
>   	case VDUSE_DEV_INJECT_CONFIG_IRQ:
> -		ret = vduse_dev_queue_irq_work(dev, &dev->inject);
> +		ret = vduse_dev_queue_irq_work(dev, &dev->inject, IRQ_UNBOUND);
>   		break;
>   	case VDUSE_VQ_SETUP: {
>   		struct vduse_vq_config config;
> @@ -1198,7 +1303,10 @@ static long vduse_dev_ioctl(struct file *file, unsigned int cmd,
>   			break;
>   
>   		index = array_index_nospec(index, dev->vq_num);
> -		ret = vduse_dev_queue_irq_work(dev, &dev->vqs[index]->inject);
> +
> +		vduse_vq_update_effective_cpu(dev->vqs[index]);
> +		ret = vduse_dev_queue_irq_work(dev, &dev->vqs[index]->inject,
> +					dev->vqs[index]->irq_effective_cpu);
>   		break;
>   	}
>   	case VDUSE_IOTLB_REG_UMEM: {
> @@ -1367,10 +1475,12 @@ static int vduse_dev_init_vqs(struct vduse_dev *dev, u32 vq_align, u32 vq_num)
>   			goto err;
>   
>   		dev->vqs[i]->index = i;
> +		dev->vqs[i]->irq_effective_cpu = -1;


IRQ_UNBOUND?

Other looks good.

Thanks


>   		INIT_WORK(&dev->vqs[i]->inject, vduse_vq_irq_inject);
>   		INIT_WORK(&dev->vqs[i]->kick, vduse_vq_kick_work);
>   		spin_lock_init(&dev->vqs[i]->kick_lock);
>   		spin_lock_init(&dev->vqs[i]->irq_lock);
> +		cpumask_setall(&dev->vqs[i]->irq_affinity);
>   	}
>   
>   	return 0;
> @@ -1858,12 +1968,15 @@ static int vduse_init(void)
>   	if (ret)
>   		goto err_cdev;
>   
> +	ret = -ENOMEM;
>   	vduse_irq_wq = alloc_workqueue("vduse-irq",
>   				WQ_HIGHPRI | WQ_SYSFS | WQ_UNBOUND, 0);
> -	if (!vduse_irq_wq) {
> -		ret = -ENOMEM;
> +	if (!vduse_irq_wq)
>   		goto err_wq;
> -	}
> +
> +	vduse_irq_bound_wq = alloc_workqueue("vduse-irq-bound", WQ_HIGHPRI, 0);
> +	if (!vduse_irq_bound_wq)
> +		goto err_bound_wq;
>   
>   	ret = vduse_domain_init();
>   	if (ret)
> @@ -1877,6 +1990,8 @@ static int vduse_init(void)
>   err_mgmtdev:
>   	vduse_domain_exit();
>   err_domain:
> +	destroy_workqueue(vduse_irq_bound_wq);
> +err_bound_wq:
>   	destroy_workqueue(vduse_irq_wq);
>   err_wq:
>   	cdev_del(&vduse_cdev);
> @@ -1896,6 +2011,7 @@ static void vduse_exit(void)
>   {
>   	vduse_mgmtdev_exit();
>   	vduse_domain_exit();
> +	destroy_workqueue(vduse_irq_bound_wq);
>   	destroy_workqueue(vduse_irq_wq);
>   	cdev_del(&vduse_cdev);
>   	device_destroy(vduse_class, vduse_major);

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 00/11] VDUSE: Improve performance

[Michael S. Tsirkin]

On Tue, Feb 28, 2023 at 05:40:59PM +0800, Xie Yongji wrote:
> Hi all,
> 
> This series introduces some ways to improve VDUSE performance.


Pls fix warnings reported by 0-day infra, dropping this for now.


> Patch 1 ~ 6 bring current interrupt affinity spreading mechanism
> to vduse device and make it possible for the virtio-blk driver
> to build the blk-mq queues based on it. This would be useful to
> mitigate the virtqueue lock contention in virtio-blk driver. In
> our test, with those patches, we could get ~50% improvement (600k
> iops -> 900k iops) when using per-cpu virtqueue.
> 
> Patch 7 adds a sysfs interface for each vduse virtqueue to change
> the affinity for IRQ callback. It would be helpful for performance
> tuning when the affinity mask contains more than one CPU.
> 
> Patch 8 ~ 9 associate an eventfd to the vdpa callback so that
> we can signal it directly during irq injection without scheduling
> an additional workqueue thread to do that.
> 
> Patch 10, 11 add a sysfs interface to support specifying bounce
> buffer size in virtio-vdpa case. The high throughput workloads
> can benefit from it. And we can also use it to reduce the memory
> overhead for small throughput workloads.
> 
> Please review, thanks!
> 
> V2 to V3:
> - Rebased to newest kernel tree
> - Export group_cpus_evenly() instead of irq_create_affinity_masks() [MST]
> - Remove the sysfs for workqueue control [Jason]
> - Associate an eventfd to the vdpa callback [Jason]
> - Signal the eventfd directly in vhost-vdpa case [Jason]
> - Use round-robin to spread IRQs between CPUs in the affinity mask [Jason]
> - Handle the cpu hotplug case on IRQ injection [Jason]
> - Remove effective IRQ affinity and balance mechanism for IRQ allocation
> 
> V1 to V2:
> - Export irq_create_affinity_masks()
> - Add set/get_vq_affinity and set_irq_affinity callbacks in vDPA
>   framework
> - Add automatic irq callback affinity support in VDUSE driver [Jason]
> - Add more backgrounds information in commit log [Jason]
> - Only support changing effective affinity when the value is a subset
>   of the IRQ callback affinity mask
> 
> Xie Yongji (11):
>   lib/group_cpus: Export group_cpus_evenly()
>   vdpa: Add set/get_vq_affinity callbacks in vdpa_config_ops
>   vdpa: Add set_irq_affinity callback in vdpa_config_ops
>   vduse: Refactor allocation for vduse virtqueues
>   vduse: Support automatic irq callback affinity
>   vduse: Support set/get_vq_affinity callbacks
>   vduse: Add sysfs interface for irq callback affinity
>   vdpa: Add eventfd for the vdpa callback
>   vduse: Signal interrupt's eventfd directly in vhost-vdpa case
>   vduse: Delay iova domain creation
>   vduse: Support specifying bounce buffer size via sysfs
> 
>  drivers/vdpa/vdpa_user/vduse_dev.c | 490 +++++++++++++++++++++++++----
>  drivers/vhost/vdpa.c               |   2 +
>  drivers/virtio/virtio_vdpa.c       |  33 ++
>  include/linux/vdpa.h               |  25 ++
>  lib/group_cpus.c                   |   1 +
>  5 files changed, 488 insertions(+), 63 deletions(-)
> 
> -- 
> 2.20.1

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 00/11] VDUSE: Improve performance

[Jason Wang]

On Fri, Mar 10, 2023 at 4:50 PM Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Tue, Feb 28, 2023 at 05:40:59PM +0800, Xie Yongji wrote:
> > Hi all,
> >
> > This series introduces some ways to improve VDUSE performance.
>
>
> Pls fix warnings reported by 0-day infra, dropping this for now.

Note that I plan to review this next week.

Thanks

>
>
> > Patch 1 ~ 6 bring current interrupt affinity spreading mechanism
> > to vduse device and make it possible for the virtio-blk driver
> > to build the blk-mq queues based on it. This would be useful to
> > mitigate the virtqueue lock contention in virtio-blk driver. In
> > our test, with those patches, we could get ~50% improvement (600k
> > iops -> 900k iops) when using per-cpu virtqueue.
> >
> > Patch 7 adds a sysfs interface for each vduse virtqueue to change
> > the affinity for IRQ callback. It would be helpful for performance
> > tuning when the affinity mask contains more than one CPU.
> >
> > Patch 8 ~ 9 associate an eventfd to the vdpa callback so that
> > we can signal it directly during irq injection without scheduling
> > an additional workqueue thread to do that.
> >
> > Patch 10, 11 add a sysfs interface to support specifying bounce
> > buffer size in virtio-vdpa case. The high throughput workloads
> > can benefit from it. And we can also use it to reduce the memory
> > overhead for small throughput workloads.
> >
> > Please review, thanks!
> >
> > V2 to V3:
> > - Rebased to newest kernel tree
> > - Export group_cpus_evenly() instead of irq_create_affinity_masks() [MST]
> > - Remove the sysfs for workqueue control [Jason]
> > - Associate an eventfd to the vdpa callback [Jason]
> > - Signal the eventfd directly in vhost-vdpa case [Jason]
> > - Use round-robin to spread IRQs between CPUs in the affinity mask [Jason]
> > - Handle the cpu hotplug case on IRQ injection [Jason]
> > - Remove effective IRQ affinity and balance mechanism for IRQ allocation
> >
> > V1 to V2:
> > - Export irq_create_affinity_masks()
> > - Add set/get_vq_affinity and set_irq_affinity callbacks in vDPA
> >   framework
> > - Add automatic irq callback affinity support in VDUSE driver [Jason]
> > - Add more backgrounds information in commit log [Jason]
> > - Only support changing effective affinity when the value is a subset
> >   of the IRQ callback affinity mask
> >
> > Xie Yongji (11):
> >   lib/group_cpus: Export group_cpus_evenly()
> >   vdpa: Add set/get_vq_affinity callbacks in vdpa_config_ops
> >   vdpa: Add set_irq_affinity callback in vdpa_config_ops
> >   vduse: Refactor allocation for vduse virtqueues
> >   vduse: Support automatic irq callback affinity
> >   vduse: Support set/get_vq_affinity callbacks
> >   vduse: Add sysfs interface for irq callback affinity
> >   vdpa: Add eventfd for the vdpa callback
> >   vduse: Signal interrupt's eventfd directly in vhost-vdpa case
> >   vduse: Delay iova domain creation
> >   vduse: Support specifying bounce buffer size via sysfs
> >
> >  drivers/vdpa/vdpa_user/vduse_dev.c | 490 +++++++++++++++++++++++++----
> >  drivers/vhost/vdpa.c               |   2 +
> >  drivers/virtio/virtio_vdpa.c       |  33 ++
> >  include/linux/vdpa.h               |  25 ++
> >  lib/group_cpus.c                   |   1 +
> >  5 files changed, 488 insertions(+), 63 deletions(-)
> >
> > --
> > 2.20.1
>

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 01/11] lib/group_cpus: Export group_cpus_evenly()

[Michael S. Tsirkin]

On Tue, Feb 28, 2023 at 05:41:00PM +0800, Xie Yongji wrote:
> Export group_cpus_evenly() so that some modules
> can make use of it to group CPUs evenly according
> to NUMA and CPU locality.
> 
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>

Thomas can I get an ack from you pls?
Anyone else?

> ---
>  lib/group_cpus.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/lib/group_cpus.c b/lib/group_cpus.c
> index 9c837a35fef7..aa3f6815bb12 100644
> --- a/lib/group_cpus.c
> +++ b/lib/group_cpus.c
> @@ -426,3 +426,4 @@ struct cpumask *group_cpus_evenly(unsigned int numgrps)
>  	return masks;
>  }
>  #endif /* CONFIG_SMP */
> +EXPORT_SYMBOL_GPL(group_cpus_evenly);
> -- 
> 2.20.1

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 01/11] lib/group_cpus: Export group_cpus_evenly()

[Jason Wang]

On Tue, Feb 28, 2023 at 5:42 PM Xie Yongji <xieyongji@bytedance.com> wrote:
>
> Export group_cpus_evenly() so that some modules
> can make use of it to group CPUs evenly according
> to NUMA and CPU locality.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>

Acked-by: Jason Wang <jasowang@redhat.com>

Thanks

> ---
>  lib/group_cpus.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/lib/group_cpus.c b/lib/group_cpus.c
> index 9c837a35fef7..aa3f6815bb12 100644
> --- a/lib/group_cpus.c
> +++ b/lib/group_cpus.c
> @@ -426,3 +426,4 @@ struct cpumask *group_cpus_evenly(unsigned int numgrps)
>         return masks;
>  }
>  #endif /* CONFIG_SMP */
> +EXPORT_SYMBOL_GPL(group_cpus_evenly);
> --
> 2.20.1
>

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 02/11] vdpa: Add set/get_vq_affinity callbacks in vdpa_config_ops

[Jason Wang]

在 2023/2/28 17:41, Xie Yongji 写道:
> This introduces set/get_vq_affinity callbacks in
> vdpa_config_ops to support interrupt affinity
> management for vdpa device drivers.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> Acked-by: Jason Wang <jasowang@redhat.com>
> ---
>   drivers/virtio/virtio_vdpa.c | 28 ++++++++++++++++++++++++++++
>   include/linux/vdpa.h         | 13 +++++++++++++
>   2 files changed, 41 insertions(+)
>
> diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c
> index d7f5af62ddaa..f72696b4c1c2 100644
> --- a/drivers/virtio/virtio_vdpa.c
> +++ b/drivers/virtio/virtio_vdpa.c
> @@ -337,6 +337,32 @@ static const char *virtio_vdpa_bus_name(struct virtio_device *vdev)
>   	return dev_name(&vdpa->dev);
>   }
>   
> +static int virtio_vdpa_set_vq_affinity(struct virtqueue *vq,
> +				       const struct cpumask *cpu_mask)
> +{
> +	struct virtio_vdpa_device *vd_dev = to_virtio_vdpa_device(vq->vdev);
> +	struct vdpa_device *vdpa = vd_dev->vdpa;
> +	const struct vdpa_config_ops *ops = vdpa->config;
> +	unsigned int index = vq->index;
> +
> +	if (ops->set_vq_affinity)
> +		return ops->set_vq_affinity(vdpa, index, cpu_mask);
> +
> +	return 0;
> +}
> +
> +static const struct cpumask *
> +virtio_vdpa_get_vq_affinity(struct virtio_device *vdev, int index)
> +{
> +	struct vdpa_device *vdpa = vd_get_vdpa(vdev);
> +	const struct vdpa_config_ops *ops = vdpa->config;
> +
> +	if (ops->get_vq_affinity)
> +		return ops->get_vq_affinity(vdpa, index);
> +
> +	return NULL;
> +}
> +
>   static const struct virtio_config_ops virtio_vdpa_config_ops = {
>   	.get		= virtio_vdpa_get,
>   	.set		= virtio_vdpa_set,
> @@ -349,6 +375,8 @@ static const struct virtio_config_ops virtio_vdpa_config_ops = {
>   	.get_features	= virtio_vdpa_get_features,
>   	.finalize_features = virtio_vdpa_finalize_features,
>   	.bus_name	= virtio_vdpa_bus_name,
> +	.set_vq_affinity = virtio_vdpa_set_vq_affinity,
> +	.get_vq_affinity = virtio_vdpa_get_vq_affinity,
>   };
>   
>   static void virtio_vdpa_release_dev(struct device *_d)
> diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h
> index 43f59ef10cc9..d61f369f9cd6 100644
> --- a/include/linux/vdpa.h
> +++ b/include/linux/vdpa.h
> @@ -250,6 +250,15 @@ struct vdpa_map_file {
>    *				@vdev: vdpa device
>    *				Returns the iova range supported by
>    *				the device.
> + * @set_vq_affinity:		Set the irq affinity of virtqueue (optional)


Nit: it's better not mention IRQ here because the virtqueue notification 
is not necessarily backed on IRQ.

Thanks


> + *				@vdev: vdpa device
> + *				@idx: virtqueue index
> + *				@cpu_mask: irq affinity mask
> + *				Returns integer: success (0) or error (< 0)
> + * @get_vq_affinity:		Get the irq affinity of virtqueue (optional)
> + *				@vdev: vdpa device
> + *				@idx: virtqueue index
> + *				Returns the irq affinity mask
>    * @set_group_asid:		Set address space identifier for a
>    *				virtqueue group (optional)
>    *				@vdev: vdpa device
> @@ -340,6 +349,10 @@ struct vdpa_config_ops {
>   			   const void *buf, unsigned int len);
>   	u32 (*get_generation)(struct vdpa_device *vdev);
>   	struct vdpa_iova_range (*get_iova_range)(struct vdpa_device *vdev);
> +	int (*set_vq_affinity)(struct vdpa_device *vdev, u16 idx,
> +			       const struct cpumask *cpu_mask);
> +	const struct cpumask *(*get_vq_affinity)(struct vdpa_device *vdev,
> +						 u16 idx);
>   
>   	/* DMA ops */
>   	int (*set_map)(struct vdpa_device *vdev, unsigned int asid,

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 03/11] vdpa: Add set_irq_affinity callback in vdpa_config_ops

[Jason Wang]

On Tue, Feb 28, 2023 at 5:42 PM Xie Yongji <xieyongji@bytedance.com> wrote:
>
> This introduces set_irq_affinity callback in
> vdpa_config_ops so that vdpa device driver can
> get the interrupt affinity hint from the virtio
> device driver. The interrupt affinity hint would
> be needed by the interrupt affinity spreading
> mechanism.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> ---
>  drivers/virtio/virtio_vdpa.c | 4 ++++
>  include/linux/vdpa.h         | 9 +++++++++
>  2 files changed, 13 insertions(+)
>
> diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c
> index f72696b4c1c2..9eee8afabda8 100644
> --- a/drivers/virtio/virtio_vdpa.c
> +++ b/drivers/virtio/virtio_vdpa.c
> @@ -282,9 +282,13 @@ static int virtio_vdpa_find_vqs(struct virtio_device *vdev, unsigned int nvqs,
>         struct virtio_vdpa_device *vd_dev = to_virtio_vdpa_device(vdev);
>         struct vdpa_device *vdpa = vd_get_vdpa(vdev);
>         const struct vdpa_config_ops *ops = vdpa->config;
> +       struct irq_affinity default_affd = { 0 };
>         struct vdpa_callback cb;
>         int i, err, queue_idx = 0;
>
> +       if (ops->set_irq_affinity)
> +               ops->set_irq_affinity(vdpa, desc ? desc : &default_affd);
> +
>         for (i = 0; i < nvqs; ++i) {
>                 if (!names[i]) {
>                         vqs[i] = NULL;
> diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h
> index d61f369f9cd6..10bd22387276 100644
> --- a/include/linux/vdpa.h
> +++ b/include/linux/vdpa.h
> @@ -259,6 +259,13 @@ struct vdpa_map_file {
>   *                             @vdev: vdpa device
>   *                             @idx: virtqueue index
>   *                             Returns the irq affinity mask
> + * @set_irq_affinity:          Pass the irq affinity hint (best effort)

Note that this could easily confuse the users. I wonder if we can
unify it with set_irq_affinity. Looking at vduse's implementation, it
should be possible.

(E.g set_vq_affinity implemented by virtio-pci are using irq affinity actually).

Thanks

> + *                             from the virtio device driver to vdpa
> + *                             driver (optional).
> + *                             Needed by the interrupt affinity spreading
> + *                             mechanism.
> + *                             @vdev: vdpa device
> + *                             @desc: irq affinity hint
>   * @set_group_asid:            Set address space identifier for a
>   *                             virtqueue group (optional)
>   *                             @vdev: vdpa device
> @@ -353,6 +360,8 @@ struct vdpa_config_ops {
>                                const struct cpumask *cpu_mask);
>         const struct cpumask *(*get_vq_affinity)(struct vdpa_device *vdev,
>                                                  u16 idx);
> +       void (*set_irq_affinity)(struct vdpa_device *vdev,
> +                                struct irq_affinity *desc);
>
>         /* DMA ops */
>         int (*set_map)(struct vdpa_device *vdev, unsigned int asid,
> --
> 2.20.1
>

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 03/11] vdpa: Add set_irq_affinity callback in vdpa_config_ops

[Jason Wang]

On Fri, Mar 17, 2023 at 3:45 PM Yongji Xie <xieyongji@bytedance.com> wrote:
>
> On Thu, Mar 16, 2023 at 12:03 PM Jason Wang <jasowang@redhat.com> wrote:
> >
> > On Tue, Feb 28, 2023 at 5:42 PM Xie Yongji <xieyongji@bytedance.com> wrote:
> > >
> > > This introduces set_irq_affinity callback in
> > > vdpa_config_ops so that vdpa device driver can
> > > get the interrupt affinity hint from the virtio
> > > device driver. The interrupt affinity hint would
> > > be needed by the interrupt affinity spreading
> > > mechanism.
> > >
> > > Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> > > ---
> > >  drivers/virtio/virtio_vdpa.c | 4 ++++
> > >  include/linux/vdpa.h         | 9 +++++++++
> > >  2 files changed, 13 insertions(+)
> > >
> > > diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c
> > > index f72696b4c1c2..9eee8afabda8 100644
> > > --- a/drivers/virtio/virtio_vdpa.c
> > > +++ b/drivers/virtio/virtio_vdpa.c
> > > @@ -282,9 +282,13 @@ static int virtio_vdpa_find_vqs(struct virtio_device *vdev, unsigned int nvqs,
> > >         struct virtio_vdpa_device *vd_dev = to_virtio_vdpa_device(vdev);
> > >         struct vdpa_device *vdpa = vd_get_vdpa(vdev);
> > >         const struct vdpa_config_ops *ops = vdpa->config;
> > > +       struct irq_affinity default_affd = { 0 };
> > >         struct vdpa_callback cb;
> > >         int i, err, queue_idx = 0;
> > >
> > > +       if (ops->set_irq_affinity)
> > > +               ops->set_irq_affinity(vdpa, desc ? desc : &default_affd);
> > > +
> > >         for (i = 0; i < nvqs; ++i) {
> > >                 if (!names[i]) {
> > >                         vqs[i] = NULL;
> > > diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h
> > > index d61f369f9cd6..10bd22387276 100644
> > > --- a/include/linux/vdpa.h
> > > +++ b/include/linux/vdpa.h
> > > @@ -259,6 +259,13 @@ struct vdpa_map_file {
> > >   *                             @vdev: vdpa device
> > >   *                             @idx: virtqueue index
> > >   *                             Returns the irq affinity mask
> > > + * @set_irq_affinity:          Pass the irq affinity hint (best effort)
> >
> > Note that this could easily confuse the users. I wonder if we can
> > unify it with set_irq_affinity. Looking at vduse's implementation, it
> > should be possible.
> >
>
> Do you mean unify set_irq_affinity() with set_vq_affinity()? Actually
> I didn't get how to achieve that. The set_vq_affinity() callback is
> called by virtio_config_ops.set_vq_affinity() but the set_irq_affinity
> is called by virtio_config_ops.find_vqs(), I don't know where to call
> the unified callback.

I meant, can we stick a single per vq affinity config ops then use
that in virtio-vpda's find_vqs() by something like:

masks = create_affinity_masks(dev->vq_num, desc);
for (i = 0; i < dev->vq_num; i++)
        config->set_vq_affinity()
...

?

Thanks

>
> Thanks,
> Yongji
>

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 08/11] vdpa: Add eventfd for the vdpa callback

[Jason Wang]

在 2023/2/28 17:41, Xie Yongji 写道:
> Add eventfd for the vdpa callback so that user
> can signal it directly instead of running the
> callback. It will be used for vhost-vdpa case.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> ---
>   drivers/vhost/vdpa.c         | 2 ++
>   drivers/virtio/virtio_vdpa.c | 1 +
>   include/linux/vdpa.h         | 3 +++
>   3 files changed, 6 insertions(+)
>
> diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c
> index dc12dbd5b43b..ae89c0ccc2bb 100644
> --- a/drivers/vhost/vdpa.c
> +++ b/drivers/vhost/vdpa.c
> @@ -599,9 +599,11 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
>   		if (vq->call_ctx.ctx) {
>   			cb.callback = vhost_vdpa_virtqueue_cb;
>   			cb.private = vq;
> +			cb.irq_ctx = vq->call_ctx.ctx;
>   		} else {
>   			cb.callback = NULL;
>   			cb.private = NULL;
> +			cb.irq_ctx = NULL;
>   		}
>   		ops->set_vq_cb(vdpa, idx, &cb);
>   		vhost_vdpa_setup_vq_irq(v, idx);
> diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c
> index 9eee8afabda8..a5cecafbc2d1 100644
> --- a/drivers/virtio/virtio_vdpa.c
> +++ b/drivers/virtio/virtio_vdpa.c
> @@ -195,6 +195,7 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index,
>   	/* Setup virtqueue callback */
>   	cb.callback = callback ? virtio_vdpa_virtqueue_cb : NULL;
>   	cb.private = info;
> +	cb.irq_ctx = NULL;
>   	ops->set_vq_cb(vdpa, index, &cb);
>   	ops->set_vq_num(vdpa, index, virtqueue_get_vring_size(vq));
>   
> diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h
> index 10bd22387276..94a7ec49583a 100644
> --- a/include/linux/vdpa.h
> +++ b/include/linux/vdpa.h
> @@ -13,10 +13,13 @@
>    * struct vdpa_calllback - vDPA callback definition.
>    * @callback: interrupt callback function
>    * @private: the data passed to the callback function
> + * @irq_ctx: the eventfd for the callback, user can signal
> + *           it directly instead of running the callback


I'd suggest to do more tweaks to mention:

1) irq_ctx is optional
2) that when the irq_ctx is set, the vDPA driver must guarantee that 
signaling it is functional equivalent to triggering the callback. When 
set, vDPA parent can signal it directly instead of triggering the callback.

>    */
>   struct vdpa_callback {
>   	irqreturn_t (*callback)(void *data);
>   	void *private;
> +	struct eventfd_ctx *irq_ctx;


There's no IRQ concept at the virtual vDPA bus level, so it's probably 
better to rename it as "trigger".

Btw, should we select EVENTFD for vDPA?

Thanks


>   };
>   
>   /**

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 08/11] vdpa: Add eventfd for the vdpa callback

[Jason Wang]

On Thu, Mar 16, 2023 at 5:25 PM Jason Wang <jasowang@redhat.com> wrote:
>
>
> 在 2023/2/28 17:41, Xie Yongji 写道:
> > Add eventfd for the vdpa callback so that user
> > can signal it directly instead of running the
> > callback. It will be used for vhost-vdpa case.
> >
> > Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> > ---
> >   drivers/vhost/vdpa.c         | 2 ++
> >   drivers/virtio/virtio_vdpa.c | 1 +
> >   include/linux/vdpa.h         | 3 +++
> >   3 files changed, 6 insertions(+)
> >
> > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c
> > index dc12dbd5b43b..ae89c0ccc2bb 100644
> > --- a/drivers/vhost/vdpa.c
> > +++ b/drivers/vhost/vdpa.c
> > @@ -599,9 +599,11 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
> >               if (vq->call_ctx.ctx) {
> >                       cb.callback = vhost_vdpa_virtqueue_cb;
> >                       cb.private = vq;
> > +                     cb.irq_ctx = vq->call_ctx.ctx;
> >               } else {
> >                       cb.callback = NULL;
> >                       cb.private = NULL;
> > +                     cb.irq_ctx = NULL;
> >               }
> >               ops->set_vq_cb(vdpa, idx, &cb);
> >               vhost_vdpa_setup_vq_irq(v, idx);
> > diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c
> > index 9eee8afabda8..a5cecafbc2d1 100644
> > --- a/drivers/virtio/virtio_vdpa.c
> > +++ b/drivers/virtio/virtio_vdpa.c
> > @@ -195,6 +195,7 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index,
> >       /* Setup virtqueue callback */
> >       cb.callback = callback ? virtio_vdpa_virtqueue_cb : NULL;
> >       cb.private = info;
> > +     cb.irq_ctx = NULL;
> >       ops->set_vq_cb(vdpa, index, &cb);
> >       ops->set_vq_num(vdpa, index, virtqueue_get_vring_size(vq));
> >
> > diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h
> > index 10bd22387276..94a7ec49583a 100644
> > --- a/include/linux/vdpa.h
> > +++ b/include/linux/vdpa.h
> > @@ -13,10 +13,13 @@
> >    * struct vdpa_calllback - vDPA callback definition.
> >    * @callback: interrupt callback function
> >    * @private: the data passed to the callback function
> > + * @irq_ctx: the eventfd for the callback, user can signal
> > + *           it directly instead of running the callback
>
>
> I'd suggest to do more tweaks to mention:
>
> 1) irq_ctx is optional
> 2) that when the irq_ctx is set, the vDPA driver must guarantee that
> signaling it is functional equivalent to triggering the callback. When
> set, vDPA parent can signal it directly instead of triggering the callback.
>
> >    */
> >   struct vdpa_callback {
> >       irqreturn_t (*callback)(void *data);
> >       void *private;
> > +     struct eventfd_ctx *irq_ctx;
>
>
> There's no IRQ concept at the virtual vDPA bus level, so it's probably
> better to rename it as "trigger".
>
> Btw, should we select EVENTFD for vDPA?

Looks like we are fine here since we only use the pointer to the eventfd_ctx.

Thanks

>
> Thanks
>
>
> >   };
> >
> >   /**

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH v3 09/11] vduse: Signal interrupt's eventfd directly in vhost-vdpa case

[Jason Wang]

On Tue, Feb 28, 2023 at 5:42 PM Xie Yongji <xieyongji@bytedance.com> wrote:
>
> Now the vdpa callback will associate an eventfd in
> vhost-vdpa case.

I'd suggest avoiding mentioning drivers since vDPA parents should not
know which vDPA driver is bound.

We could say "signal vq trigger eventfd directly if possible"?

With those tweaked.

Acked-by: Jason Wang <jasowang@redhat.com>

Thanks

> For performance reasons, VDUSE can
> signal it directly during irq injection.
>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> ---
>  drivers/vdpa/vdpa_user/vduse_dev.c | 27 +++++++++++++++++++++++----
>  1 file changed, 23 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/vdpa/vdpa_user/vduse_dev.c b/drivers/vdpa/vdpa_user/vduse_dev.c
> index 869cc7860d82..56f3c2480c2a 100644
> --- a/drivers/vdpa/vdpa_user/vduse_dev.c
> +++ b/drivers/vdpa/vdpa_user/vduse_dev.c
> @@ -461,6 +461,7 @@ static void vduse_dev_reset(struct vduse_dev *dev)
>                 spin_lock(&vq->irq_lock);
>                 vq->cb.callback = NULL;
>                 vq->cb.private = NULL;
> +               vq->cb.irq_ctx = NULL;
>                 spin_unlock(&vq->irq_lock);
>                 flush_work(&vq->inject);
>                 flush_work(&vq->kick);
> @@ -526,6 +527,7 @@ static void vduse_vdpa_set_vq_cb(struct vdpa_device *vdpa, u16 idx,
>         spin_lock(&vq->irq_lock);
>         vq->cb.callback = cb->callback;
>         vq->cb.private = cb->private;
> +       vq->cb.irq_ctx = cb->irq_ctx;
>         spin_unlock(&vq->irq_lock);
>  }
>
> @@ -1020,6 +1022,20 @@ static void vduse_vq_irq_inject(struct work_struct *work)
>         spin_unlock_irq(&vq->irq_lock);
>  }
>
> +static bool vduse_vq_signal_irqfd(struct vduse_virtqueue *vq)
> +{
> +       bool signal = false;
> +
> +       spin_lock_irq(&vq->irq_lock);
> +       if (vq->ready && vq->cb.irq_ctx) {
> +               eventfd_signal(vq->cb.irq_ctx, 1);
> +               signal = true;
> +       }
> +       spin_unlock_irq(&vq->irq_lock);
> +
> +       return signal;
> +}
> +
>  static int vduse_dev_queue_irq_work(struct vduse_dev *dev,
>                                     struct work_struct *irq_work,
>                                     int irq_effective_cpu)
> @@ -1322,11 +1338,14 @@ static long vduse_dev_ioctl(struct file *file, unsigned int cmd,
>                 if (index >= dev->vq_num)
>                         break;
>
> +               ret = 0;
>                 index = array_index_nospec(index, dev->vq_num);
> -
> -               vduse_vq_update_effective_cpu(dev->vqs[index]);
> -               ret = vduse_dev_queue_irq_work(dev, &dev->vqs[index]->inject,
> -                                       dev->vqs[index]->irq_effective_cpu);
> +               if (!vduse_vq_signal_irqfd(dev->vqs[index])) {
> +                       vduse_vq_update_effective_cpu(dev->vqs[index]);
> +                       ret = vduse_dev_queue_irq_work(dev,
> +                                               &dev->vqs[index]->inject,
> +                                               dev->vqs[index]->irq_effective_cpu);
> +               }
>                 break;
>         }
>         case VDUSE_IOTLB_REG_UMEM: {
> --
> 2.20.1
>

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization