Re: [libvirt] [PATCH 20/21] qemu_hotplug: consolidate all common detach code in qemuDomainDetachDeviceLive

On Fri, Mar 22, 2019 at 10:08:17 -0400, Laine Stump wrote: > On 3/22/19 8:32 AM, Peter Krempa wrote: >> On Thu, Mar 21, 2019 at 18:29:00 -0400, Laine Stump wrote: >>> Now that all the qemuDomainDetachPrep*() functions look nearly >>> identical at the end, we can put one copy of that identical code in >>> qemuDomainDetachDeviceLive() at the point after the individual prep >>> functions have been called, and remove the duplicated code from all >>> the prep functions. The code to locate the target "detach" device >>> based on the "match" device remains, as do all device-type-specific >>> validations. >>> >>> Unfortunately there are a few things going on at once in this patch, >>> which makes it a bit more difficult to follow than the others; it was >>> just impossible to do the changes in stages and still have a >>> buildable/testable tree at each step. >>> >>> The other changes of note: >>> >>> * The individual prep functions no longer need their driver or async >>> args, so those are removed, as are the local "ret" variables, since >>> in all cases the functions just directly return -1 or 0. >>> >>> * Some of the prep functions were checking for a valid alias and/or >>> for attempts to detach a multifunction PCI device, but not all. In >>> fact, both checks are valid (or at least harmless) for *all* device >>> types, so they are removed from the prep functions, and done a >>> single time in the common function. >>> >>> (any attempts to *create* an alias when there isn't one has been >>> removed, since that is doomed to failure anyway; the only way the >>> device wouldn't have an alias is if 1) the domain was created by >>> calling virsh qemu-attach to attach an existing qemu process to >>> libvirt, and 2) the qemu command that started said process used "old >>> style" arguments for creating devices that didn't have any device >>> ids. Even if we constructed a device id for one of these devices, >>> qemu wouldn't recognize it in the device_del command anyway, so we >>> may as well fail earlier with "device missing alias" rather than >>> failing later with "couldn't delete device net0".) >>> >>> * Only one type of device has shutdown code that must not be called >>> until after *all* validation of the device is done (including >> Why don't we put that to the RemoveDevice handler for network? We >> teardown all related stuff usually after unplug of the frontend. > > That is a good question! I'm not sure why network devices have this part of > their shutdown handled early. I don't think we should make that change here > though. I can insert an earlier patch to do it, hoping that it won't break > anything, or I can make a later patch that moves it after this change, which > seems kind of superfluous, but would make it simpler to revert in case it > causes a regression. What's your opinion? I honestly have no idea what that cleanup is done for network devices, it just feels wrong to teardown stuff prior to the OS even giving up the device.

Changing it prior will result in cleaner code, fixing it after will be easily revertable. I'd probably go for the latter.

> > >>> checking for multifunction PCI and valid alias, which is done in the >>> toplevel common code). For this reason, the Net function has been >>> split in two, with the 2nd half (qemuDomainDetachShutdownNet()) >>> called from the common function, right before sending the delete >>> command to qemu. >>> >>> Signed-off-by: Laine Stump <laine(a)laine.org&gt; >>> --- >>> src/qemu/qemu_hotplug.c | 497 ++++++++++++---------------------------- >>> 1 file changed, 149 insertions(+), 348 deletions(-) >>> >>> diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c >>> index 5e5ffe16d3..de7a7a2c95 100644 >>> --- a/src/qemu/qemu_hotplug.c >>> +++ b/src/qemu/qemu_hotplug.c >>> @@ -5208,7 +5208,7 @@ qemuDomainRemoveRedirdevDevice(virQEMUDriverPtr driver, >>> } >>> -static inline void >>> +static void >> Use ATTRIBUTE_UNUSED rather than inline. > > I couldn't remember where was the proper place to put ATTRIBUTE_UNUSED in > the function definition, and no matter where I put it it didn't work, so I > gave up and used inline. Obviously that doesn't work for all compilers > though, so I'll figure out the right place. static void ATTRIBUTE_UNUSED virAsdf(int qwer) { }