[libvirt PATCH 0/7] remove support for unpriv_sgio

Andrea Bolognani abologna at redhat.com
Wed Jan 19 10:10:38 UTC 2022


On Thu, Jan 13, 2022 at 06:47:35PM +0100, Ján Tomko wrote:
> This feature was never merged in upstream kernel and only made it
> to upstream libvirt by mistakes.
>
> Remove it as well as the 'sharedDevices' hash table that was
> only used for this purpose.
>
> The 'sgio' attribute is preserved for compatibility reasons
> and its 'filtered' value is left intact.
>
> An error is added to the XML validator for the 'unfiltered'
> value, which means 'unpriv_sgio'
>
> Ján Tomko (7):
>   util: remove virGetUnprivSGIOSysfsPath
>   conf: reject unfiltered sgio on validation
>   qemu: remove all use of SGIO
>   qemu: remove sharedDevices hash table
>   util: remove virGetDeviceID
>   util: remove {Get,Set}UnprivSGIO
>   qemu: simplify qemuProcessSetupRawIO
>
>  src/conf/domain_validate.c                    |  11 +
>  src/libvirt_private.syms                      |   4 -
>  src/qemu/qemu_conf.c                          | 448 ------------------
>  src/qemu/qemu_conf.h                          |  37 --
>  src/qemu/qemu_driver.c                        |   3 -
>  src/qemu/qemu_hostdev.c                       |  31 --
>  src/qemu/qemu_hotplug.c                       |  23 -
>  src/qemu/qemu_process.c                       |  36 +-
>  src/util/virutil.c                            | 122 -----
>  src/util/virutil.h                            |  12 -
>  tests/qemuhotplugmock.c                       |  17 -
>  .../disk-scsi-lun-passthrough-sgio.xml        |  35 --
>  tests/qemuxml2argvdata/hostdev-scsi-rawio.xml |  37 --
>  tests/qemuxml2argvdata/hostdev-scsi-sgio.xml  |  37 --
>  .../disk-scsi-lun-passthrough-sgio.xml        |  46 --
>  .../qemuxml2xmloutdata/hostdev-scsi-rawio.xml |  47 --
>  .../qemuxml2xmloutdata/hostdev-scsi-sgio.xml  |  47 --
>  tests/qemuxml2xmltest.c                       |  11 -
>  18 files changed, 14 insertions(+), 990 deletions(-)

Stunning diffstat. I feel like I might cry.

I left a couple of small comments, but either way

  Reviewed-by: Andrea Bolognani <abologna at redhat.com>

-- 
Andrea Bolognani / Red Hat / Virtualization





More information about the libvir-list mailing list