| Message ID | 5c8fb8141c0f92ccaa78cdc169b8544f7634ce65.1563325215.git-series.marmarek@invisiblethingslab.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Fix PCI passthrough for HVM with stubdomain | expand |
On Wed, Jul 17, 2019 at 03:00:44AM +0200, Marek Marczykowski-Górecki wrote: > Add libxc wrapper for PHYSDEVOP_msi_control introduced in previous > commit. > > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> LGTM, albeit I find the usage of int instead of unsigned int for the SBDF kind of weird, but it's inline with the other functions, so I guess there's a reason for it? I assume this will be used by an upcoming QEMU patch? Thanks, Roger.
On Wed, Jul 17, 2019 at 12:21:58PM +0200, Roger Pau Monné wrote: > On Wed, Jul 17, 2019 at 03:00:44AM +0200, Marek Marczykowski-Górecki wrote: > > Add libxc wrapper for PHYSDEVOP_msi_control introduced in previous > > commit. > > > > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> > > LGTM, albeit I find the usage of int instead of unsigned int for the > SBDF kind of weird, but it's inline with the other functions, so I > guess there's a reason for it? Yes, it was based on looking at other places. But I don't know if there is any specific reason for it. > I assume this will be used by an upcoming QEMU patch? Yes.
On Thu, Jul 18, 2019 at 02:12:20AM +0200, Marek Marczykowski-Górecki wrote: > On Wed, Jul 17, 2019 at 12:21:58PM +0200, Roger Pau Monné wrote: > > On Wed, Jul 17, 2019 at 03:00:44AM +0200, Marek Marczykowski-Górecki wrote: > > > Add libxc wrapper for PHYSDEVOP_msi_control introduced in previous > > > commit. > > > > > > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> > > > > LGTM, albeit I find the usage of int instead of unsigned int for the > > SBDF kind of weird, but it's inline with the other functions, so I > > guess there's a reason for it? > > Yes, it was based on looking at other places. But I don't know if there > is any specific reason for it. Anyway, since using unsigned or signed is not really that relevant here, and seeing how other functions are defined: Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> With just a couple of nits: you don't actually need rc, you can just `return do_physdev_op...`, and you could also initialize physdev_msi_control at declaration, but I don't have a strong opinion on any of those, so you can keep the RB regardless. Thanks!
diff --git a/tools/libxc/include/xenctrl.h b/tools/libxc/include/xenctrl.h index 538007a..826d10d 100644 --- a/tools/libxc/include/xenctrl.h +++ b/tools/libxc/include/xenctrl.h @@ -1638,6 +1638,12 @@ int xc_physdev_unmap_pirq(xc_interface *xch, uint32_t domid, int pirq); +int xc_physdev_msi_control(xc_interface *xch, + int seg, + int bus, + int devfn, + int flags); + /* * LOGGING AND ERROR REPORTING */ diff --git a/tools/libxc/xc_physdev.c b/tools/libxc/xc_physdev.c index 460a8e7..a25a117 100644 --- a/tools/libxc/xc_physdev.c +++ b/tools/libxc/xc_physdev.c @@ -111,3 +111,22 @@ int xc_physdev_unmap_pirq(xc_interface *xch, return rc; } +int xc_physdev_msi_control(xc_interface *xch, + int seg, + int bus, + int devfn, + int flags) +{ + int rc; + struct physdev_msi_control op; + + memset(&op, 0, sizeof(struct physdev_msi_control)); + op.seg = seg; + op.bus = bus; + op.devfn = devfn; + op.flags = flags; + + rc = do_physdev_op(xch, PHYSDEVOP_msi_control, &op, sizeof(op)); + + return rc; +}
Add libxc wrapper for PHYSDEVOP_msi_control introduced in previous commit. Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> --- Changes in v3: - new patch Changes in v4: - adjust for updated previous patch Changes in v5: - rename to PHYSDEVOP_msi_control, adjust arguments --- tools/libxc/include/xenctrl.h | 6 ++++++ tools/libxc/xc_physdev.c | 19 +++++++++++++++++++ 2 files changed, 25 insertions(+)