qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH 3/3] vfio/pci: pass an error object to vfio_msix


From: Auger Eric
Subject: Re: [Qemu-devel] [PATCH 3/3] vfio/pci: pass an error object to vfio_msix_early_setup
Date: Mon, 12 Sep 2016 17:15:40 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.1.1

Hi Markus,

On 12/09/2016 14:58, Markus Armbruster wrote:
> Eric Auger <address@hidden> writes:
> 
>> Let's expand the usage of QEMU Error objects to vfio_msix_early_setup.
>>
>> Signed-off-by: Eric Auger <address@hidden>
>> ---
>>  hw/vfio/pci.c | 28 ++++++++++++++++------------
>>  1 file changed, 16 insertions(+), 12 deletions(-)
>>
>> diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
>> index f7768e9..b132bf1 100644
>> --- a/hw/vfio/pci.c
>> +++ b/hw/vfio/pci.c
>> @@ -1275,7 +1275,7 @@ static void vfio_pci_fixup_msix_region(VFIOPCIDevice 
>> *vdev)
>>   * need to first look for where the MSI-X table lives.  So we
>>   * unfortunately split MSI-X setup across two functions.
>>   */
>> -static int vfio_msix_early_setup(VFIOPCIDevice *vdev)
>> +static void vfio_msix_early_setup(VFIOPCIDevice *vdev, Error **errp)
>>  {
>>      uint8_t pos;
>>      uint16_t ctrl;
>> @@ -1285,22 +1285,25 @@ static int vfio_msix_early_setup(VFIOPCIDevice *vdev)
>>  
>>      pos = pci_find_capability(&vdev->pdev, PCI_CAP_ID_MSIX);
>>      if (!pos) {
>> -        return 0;
>> +        return;
>>      }
>>  
>>      if (pread(fd, &ctrl, sizeof(ctrl),
>>                vdev->config_offset + pos + PCI_MSIX_FLAGS) != sizeof(ctrl)) {
>> -        return -errno;
>> +        error_setg_errno(errp, -errno, "failed to read PCI MSIX FLAGS");
>> +        return;
> 
> Looks like a bug fix to me: before your series, this is a silent
> failure.  PATCH 1 makes it non-silent, with a generic error message, and
> this patch replaces it with a specific error message.
Actually the init process was stopped without reporting the actual
issue. But that's the goal of the series. I will improve the commit
message.
> 
>>      }
>>  
>>      if (pread(fd, &table, sizeof(table),
>>                vdev->config_offset + pos + PCI_MSIX_TABLE) != sizeof(table)) 
>> {
>> -        return -errno;
>> +        error_setg_errno(errp, -errno, "failed to read PCI MSIX TABLE");
>> +        return;
> 
> Likewise.
> 
>>      }
>>  
>>      if (pread(fd, &pba, sizeof(pba),
>>                vdev->config_offset + pos + PCI_MSIX_PBA) != sizeof(pba)) {
>> -        return -errno;
>> +        error_setg_errno(errp, -errno, "failed to read PCI MSIX PBA");
>> +        return;
> 
> Likewise.
> 
>>      }
>>  
>>      ctrl = le16_to_cpu(ctrl);
>> @@ -1330,10 +1333,11 @@ static int vfio_msix_early_setup(VFIOPCIDevice *vdev)
>>              (vdev->device_id & 0xff00) == 0x5800) {
>>              msix->pba_offset = 0x1000;
>>          } else {
>> -            error_report("vfio: Hardware reports invalid configuration, "
>> -                         "MSIX PBA outside of specified BAR");
>> +            error_setg_errno(errp, -EINVAL,
>> +                       "hardware reports invalid configuration, "
>> +                       "MSIX PBA outside of specified BAR");
>>              g_free(msix);
>> -            return -EINVAL;
>> +            return;
>>          }
>>      }
>>  
>> @@ -1343,7 +1347,7 @@ static int vfio_msix_early_setup(VFIOPCIDevice *vdev)
>>  
>>      vfio_pci_fixup_msix_region(vdev);
>>  
>> -    return 0;
>> +    return;
>>  }
>>  
>>  static int vfio_msix_setup(VFIOPCIDevice *vdev, int pos)
>> @@ -2642,9 +2646,9 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
>>  
>>      vfio_pci_size_rom(vdev);
>>  
>> -    ret = vfio_msix_early_setup(vdev);
>> -    if (ret) {
>> -        error_setg_errno(errp, ret, "msix early setup failure");
>> +    vfio_msix_early_setup(vdev, &err);
>> +    if (err) {
>> +        error_propagate(errp, err);
>>          goto error;
>>      }
> 
> vfio_realize() still calls a few functions that report errors with
> error_report(), including vfio_get_group(), vfio_get_device(),
> vfio_pci_igd_opregion_init(), vfio_intx_enable().  If you prefer not to
> complete the job right now, please mark such calls FIXME.
I will complete the job but I just wanted to get some first feedbacks.
Hopefully you already pointed out some issues!

Thank you for your time.

Best Regards

Eric
> 



reply via email to

[Prev in Thread] Current Thread [Next in Thread]