On 10/21/2015 06:44 PM, John Ferlan wrote:
On 10/20/2015 11:20 AM, Andrea Bolognani wrote:
No functional changes.
---
src/qemu/qemu_hostdev.c | 40 ++++++++++++++++++++--------------------
src/qemu/qemu_hostdev.h | 22 +++++++++++-----------
src/qemu/qemu_hotplug.c | 9 ++++-----
3 files changed, 35 insertions(+), 36 deletions(-)
diff --git a/src/qemu/qemu_hostdev.c b/src/qemu/qemu_hostdev.c
index 706db0c..a4409d6 100644
--- a/src/qemu/qemu_hostdev.c
+++ b/src/qemu/qemu_hostdev.c
@@ -156,9 +156,9 @@ qemuHostdevHostSupportsPassthroughLegacy(void)
static bool
-qemuPrepareHostdevPCICheckSupport(virDomainHostdevDefPtr *hostdevs,
- size_t nhostdevs,
- virQEMUCapsPtr qemuCaps)
+qemuPrepareHostPCIDevicesCheckSupport(virDomainHostdevDefPtr *hostdevs,
+ size_t nhostdevs,
+ virQEMUCapsPtr qemuCaps)
{
bool supportsPassthroughKVM = qemuHostdevHostSupportsPassthroughLegacy();
bool supportsPassthroughVFIO = qemuHostdevHostSupportsPassthroughVFIO();
@@ -219,18 +219,18 @@ qemuPrepareHostdevPCICheckSupport(virDomainHostdevDefPtr *hostdevs,
}
int
-qemuPrepareHostdevPCIDevices(virQEMUDriverPtr driver,
- const char *name,
- const unsigned char *uuid,
- virDomainHostdevDefPtr *hostdevs,
- int nhostdevs,
- virQEMUCapsPtr qemuCaps,
- unsigned int flags)
+qemuPrepareHostPCIDevices(virQEMUDriverPtr driver,
+ const char *name,
+ const unsigned char *uuid,
+ virDomainHostdevDefPtr *hostdevs,
+ int nhostdevs,
+ virQEMUCapsPtr qemuCaps,
+ unsigned int flags)
{
int ret = -1;
virHostdevManagerPtr hostdev_mgr = driver->hostdevMgr;
- if (!qemuPrepareHostdevPCICheckSupport(hostdevs, nhostdevs, qemuCaps))
+ if (!qemuPrepareHostPCIDevicesCheckSupport(hostdevs, nhostdevs, qemuCaps))
goto out;
ret = virHostdevPreparePCIDevices(hostdev_mgr, QEMU_DRIVER_NAME,
Typically when searching I can change the "virDomainFoo" type string to
a "qemuDomainFoo" string and find the qemu API. Here those you'll see
there's a virHostDevPreparePCIDevices(), so in a way I'd expect a
qemuHostdevPreparePCIDevices
Since is a "qemu_hostdev.{c,h}" type change, then theoretically all the
API's in here should be qemuHostdevFoo, right?
Rather than qemuPrepareFoo, qemuUpdateFoo, and qemuDomainReAttachFoo
It'd be qemuHostdevPrepare, qemuHostdevUpdate, and qemuHostdevReAttach
and your new API would be qemuHostdevUpdateActiveDevices
I had a similar reply queued up, but got interrupted by dinner:
You've changes all Hostdev into Host, but in all of these cases they
really are <hostdev> devices, so I think the function name is clearer if
it has Hostdev in it rather than Host.
Also, if we're going to normalize the names, it might be better to try
and fit in with the naming convention that we've tried to start using in
other places - virModuleNameModifierAction or
virModuleNameActionModifier (or in the case of functions in the qemu
directory qemuModuleNameModifierAction() or
qemuModuleNameActionModifier()), for example virNetdevMacVLanCreate().
So instead of qemuPrepareHostdevPCIDevices() it could maybe be
qemuHostdevPCIPrepareDevices(), qemuHostdevUSBPrepareDevices() and
qemuHostdevSCSIPrepareDevices() (or maybe qemuHostdev*DevicesPrepare(),
depending on how anal you want to be about putting the verb at the end)
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list