qemu-devel
[Top][All Lists]
Advanced

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

Re: [PULL v3 23/32] s390x/pci: Add routine to get the vfio dma available


From: Philippe Mathieu-Daudé
Subject: Re: [PULL v3 23/32] s390x/pci: Add routine to get the vfio dma available count
Date: Tue, 3 Nov 2020 11:57:47 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.3.1

On 11/3/20 11:45 AM, Cornelia Huck wrote:
> On Tue, 3 Nov 2020 06:49:13 +0100
> Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
> 
>> Hi Matthew,
>>
>> On 11/1/20 10:02 PM, Alex Williamson wrote:
>>> From: Matthew Rosato <mjrosato@linux.ibm.com>
>>>
>>> Create new files for separating out vfio-specific work for s390
>>> pci. Add the first such routine, which issues VFIO_IOMMU_GET_INFO
>>> ioctl to collect the current dma available count.
>>>
>>> Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
>>> Reviewed-by: Cornelia Huck <cohuck@redhat.com>
>>> [aw: Fix non-Linux build with CONFIG_LINUX]
>>> Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
>>> ---
>>>  hw/s390x/meson.build             |    1 +
>>>  hw/s390x/s390-pci-vfio.c         |   54 
>>> ++++++++++++++++++++++++++++++++++++++
>>>  include/hw/s390x/s390-pci-vfio.h |   24 +++++++++++++++++
>>>  3 files changed, 79 insertions(+)
>>>  create mode 100644 hw/s390x/s390-pci-vfio.c
>>>  create mode 100644 include/hw/s390x/s390-pci-vfio.h
>>>
>>> diff --git a/hw/s390x/meson.build b/hw/s390x/meson.build
>>> index 948ceae7a7b3..f4663a835514 100644
>>> --- a/hw/s390x/meson.build
>>> +++ b/hw/s390x/meson.build
>>> @@ -27,6 +27,7 @@ s390x_ss.add(when: 'CONFIG_KVM', if_true: files(
>>>  ))
>>>  s390x_ss.add(when: 'CONFIG_S390_CCW_VIRTIO', if_true: 
>>> files('s390-virtio-ccw.c'))
>>>  s390x_ss.add(when: 'CONFIG_TERMINAL3270', if_true: files('3270-ccw.c'))
>>> +s390x_ss.add(when: 'CONFIG_LINUX', if_true: files('s390-pci-vfio.c'))
> 
> I think we need s/CONFIG_LINUX/CONFIG_VFIO/ here.
> 
>>>  
>>>  virtio_ss = ss.source_set()
>>>  virtio_ss.add(files('virtio-ccw.c'))
>>> diff --git a/hw/s390x/s390-pci-vfio.c b/hw/s390x/s390-pci-vfio.c
>>> new file mode 100644
>>> index 000000000000..cb3f4d98adf8
>>> --- /dev/null
>>> +++ b/hw/s390x/s390-pci-vfio.c
>>> @@ -0,0 +1,54 @@
>>> +/*
>>> + * s390 vfio-pci interfaces
>>> + *
>>> + * Copyright 2020 IBM Corp.
>>> + * Author(s): Matthew Rosato <mjrosato@linux.ibm.com>
>>> + *
>>> + * This work is licensed under the terms of the GNU GPL, version 2 or (at
>>> + * your option) any later version. See the COPYING file in the top-level
>>> + * directory.
>>> + */
>>> +
>>> +#include <sys/ioctl.h>
>>> +
>>> +#include "qemu/osdep.h"
>>> +#include "hw/s390x/s390-pci-vfio.h"
>>> +#include "hw/vfio/vfio-common.h"
>>> +
>>> +/*
>>> + * Get the current DMA available count from vfio.  Returns true if vfio is
>>> + * limiting DMA requests, false otherwise.  The current available count 
>>> read
>>> + * from vfio is returned in avail.
>>> + */
>>> +bool s390_pci_update_dma_avail(int fd, unsigned int *avail)
>>> +{
>>> +    g_autofree struct vfio_iommu_type1_info *info;
>>> +    uint32_t argsz;
>>> +
>>> +    assert(avail);
>>> +
>>> +    argsz = sizeof(struct vfio_iommu_type1_info);
>>> +    info = g_malloc0(argsz);
>>> +
>>> +    /*
>>> +     * If the specified argsz is not large enough to contain all 
>>> capabilities
>>> +     * it will be updated upon return from the ioctl.  Retry until we have
>>> +     * a big enough buffer to hold the entire capability chain.
>>> +     */
>>> +retry:
>>> +    info->argsz = argsz;
>>> +
>>> +    if (ioctl(fd, VFIO_IOMMU_GET_INFO, info)) {
>>> +        return false;
>>> +    }
>>> +
>>> +    if (info->argsz > argsz) {
>>> +        argsz = info->argsz;
>>> +        info = g_realloc(info, argsz);
>>> +        goto retry;
>>> +    }
>>> +
>>> +    /* If the capability exists, update with the current value */
>>> +    return vfio_get_info_dma_avail(info, avail);
>>> +}  
>>
>> --without-default-devices configuration is broken, bisected
>> to this commit:
> 
> Hm. Is this run in any of the CI tests? I did not even realize it
> existed :)

Paolo added it to Travis-CI 22 months ago:

commit 20885b5b1697247038018f3a7dfe8a84adf1b8e9
Author: Paolo Bonzini <pbonzini@redhat.com>
Date:   Thu Jan 31 15:57:53 2019 +0100

    .travis.yml: test that no-default-device builds do not regress

    Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

diff --git a/.travis.yml b/.travis.yml
index cca57f43148..e942175dd30 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -149,6 +149,12 @@ matrix:
         - ${SRC_DIR}/scripts/travis/coverage-summary.sh


+    # We manually include builds which we disable "make check" for
+    - env:
+        - CONFIG="--without-default-devices"
+        - TEST_CMD=""
+
+
     # We manually include builds which we disable "make check" for
     - env:
         - CONFIG="--enable-debug --enable-tcg-interpreter"
---

> 
>>
>> hw/s390x/s390-pci-vfio.c:52: undefined reference to
>> `vfio_get_info_dma_avail'
>>
>> Can you have a look please?
>>
>> Thanks.
>>
> 




reply via email to

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