Re: [Qemu-devel] vhost-user: broken mem regions

[Nikolay Nikolaev]

On Thu, Jun 26, 2014 at 12:37 AM, Damjan Marion (damarion)
<address@hidden> wrote:
>
> On 25 Jun 2014, at 20:18, Michael S. Tsirkin <address@hidden> wrote:
>
>> On Wed, Jun 25, 2014 at 09:13:36PM +0300, Nikolay Nikolaev wrote:
>>> On Wed, Jun 25, 2014 at 9:07 PM, Michael S. Tsirkin <address@hidden> wrote:
>>>> On Wed, Jun 25, 2014 at 07:56:46PM +0300, Nikolay Nikolaev wrote:
>>>>> On Wed, Jun 25, 2014 at 7:44 PM, Paolo Bonzini <address@hidden> wrote:
>>>>>>> nregions: 4
>>>>>>> region:
>>>>>>>      gpa = 0x100000000
>>>>>>>      size = 3221225472
>>>>>>>      ua = 0x2aab6ac00000
>>>>>>
>>>>>> High memory, above 3 gigabytes.
>>>>>>
>>>>>>> region:
>>>>>>>      gpa = 0xFFFC0000
>>>>>>>      size = 262144
>>>>>>>      ua = 0x7fc13d200000
>>>>>>
>>>>>> This is the BIOS.  There shouldn't be any FD for this one, it
>>>>>> is not allocated in hugetlbfs.
>>>>>>
>>>>>>> region:
>>>>>>>      gpa = 0x0
>>>>>>>      size = 655360
>>>>>>>      ua = 0x2aaaaac00000
>>>>>>> region:
>>>>>>>      gpa = 0xC0000
>>>>>>>      size = 3220439040
>>>>>>>      ua = 0x2aaaaacc0000
>>>>>>
>>>>>> Together, it's the first 3 GB of memory.
>>>>>>
>>>>>> I understand now what you mean.  Yeah, the format should be changed
>>>>>> to include the offset (why does vhost-user need the ua at all?
>>>>> The vring addresses are QEMU UA addresses. Of course vhost-user can
>>>>> translate them to guest physical before sending the message.
>>>>
>>>> It seems useful to have them as QEMU UA, this will allow
>>>> frontends other than virtio where QEMU operates the
>>>> ring.
>>>
>>> OK - so there will be a new offset field. That's fine with me. What
>>> would be the deadline for such change?
>>> It's not exactly bugfix. On the other hand there's no wide adoption of
>>> the protocol so it's still not critical to change it.
>>
>> Right. So you have to do this before the hard freeze.
>> And don't cut it too fine either :)
>
> OK, i did the change. Pasting code here for review, to keep the thread.
>
> diff --git a/exec.c b/exec.c
> index c849405..a94c583 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -1456,6 +1456,13 @@ int qemu_get_ram_fd(ram_addr_t addr)
>      return block->fd;
>  }
>
> +void *qemu_get_ram_block_host_ptr(ram_addr_t addr)
> +{
> +    RAMBlock *block = qemu_get_ram_block(addr);
> +
> +    return block->host;
> +}
> +
>  /* Return a host pointer to ram allocated with qemu_ram_alloc.
>     With the exception of the softmmu code in this file, this should
>     only be used for local memory (e.g. video ram) that the device owns,
> diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c
> index 0df6a93..0cef2d3 100644
> --- a/hw/virtio/vhost-user.c
> +++ b/hw/virtio/vhost-user.c
> @@ -14,6 +14,7 @@
>  #include "sysemu/kvm.h"
>  #include "qemu/error-report.h"
>  #include "qemu/sockets.h"
> +#include "exec/ram_addr.h"
>
>  #include <fcntl.h>
>  #include <unistd.h>
> @@ -47,6 +48,7 @@ typedef struct VhostUserMemoryRegion {
>      uint64_t guest_phys_addr;
>      uint64_t memory_size;
>      uint64_t userspace_addr;
> +    uint64_t shm_offset;
It's not bound to shm. Can it be mmap_offset?
>  } VhostUserMemoryRegion;
>
>  typedef struct VhostUserMemory {
> @@ -183,10 +185,10 @@ static int vhost_user_call(struct vhost_dev *dev, 
> unsigned long int request,
>  {
>      VhostUserMsg msg;
>      VhostUserRequest msg_request;
> -    RAMBlock *block = 0;
>      struct vhost_vring_file *file = 0;
>      int need_reply = 0;
>      int fds[VHOST_MEMORY_MAX_NREGIONS];
> +    int i, fd;
>      size_t fd_num = 0;
>
>      assert(dev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER);
> @@ -212,14 +214,16 @@ static int vhost_user_call(struct vhost_dev *dev, 
> unsigned long int request,
>          break;
>
>      case VHOST_SET_MEM_TABLE:
> -        QTAILQ_FOREACH(block, &ram_list.blocks, next)
> -        {
> -            if (block->fd > 0) {
> -                msg.memory.regions[fd_num].userspace_addr =
> -                    (uintptr_t) block->host;
> -                msg.memory.regions[fd_num].memory_size = block->length;
> -                msg.memory.regions[fd_num].guest_phys_addr = block->offset;
> -                fds[fd_num++] = block->fd;
> +        for (i = 0; i < dev->mem->nregions; ++i) {
> +            struct vhost_memory_region *reg = dev->mem->regions + i;
> +            fd = qemu_get_ram_fd(reg->guest_phys_addr);
> +            if (fd > 0) {
> +                msg.memory.regions[fd_num].userspace_addr = 
> reg->userspace_addr;
> +                msg.memory.regions[fd_num].memory_size  = reg->memory_size;
> +                msg.memory.regions[fd_num].guest_phys_addr = 
> reg->guest_phys_addr;
> +                msg.memory.regions[fd_num].shm_offset = reg->userspace_addr -
> +                    (ram_addr_t) 
> qemu_get_ram_block_host_ptr(reg->guest_phys_addr);
It may be a good idea to assert that fd_num is less than
VHOST_MEMORY_MAX_NREGIONS
> +                fds[fd_num++] = fd;
>              }
>          }
>
> diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h
> index 55ca676..e9eb831 100644
> --- a/include/exec/ram_addr.h
> +++ b/include/exec/ram_addr.h
> @@ -29,6 +29,7 @@ ram_addr_t qemu_ram_alloc_from_ptr(ram_addr_t size, void 
> *host,
>                                     MemoryRegion *mr);
>  ram_addr_t qemu_ram_alloc(ram_addr_t size, MemoryRegion *mr);
>  int qemu_get_ram_fd(ram_addr_t addr);
> +void *qemu_get_ram_block_host_ptr(ram_addr_t addr);
>  void *qemu_get_ram_ptr(ram_addr_t addr);
>  void qemu_ram_free(ram_addr_t addr);
>  void qemu_ram_free_from_ptr(ram_addr_t addr);
>
>
>

In general proposed changes are OK with snabbswitch vhost-user
implementation. We'll adapt once this is upstream.

regards,
Nikolay Nikolaev