It's essentially dead code. The only way hostdev->shareable
can be true is during virDomainHostdevDefParseXML when the
result of virHostdevIsSCSIDevice is true.
Signed-off-by: John Ferlan <jferlan@redhat.com>
---
src/qemu/qemu_conf.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c
index 20952e9607..945725566c 100644
--- a/src/qemu/qemu_conf.c
+++ b/src/qemu/qemu_conf.c
@@ -1438,8 +1438,7 @@ static bool
qemuIsSharedHostdev(virDomainHostdevDefPtr hostdev)
{
return (hostdev->shareable &&
- (virHostdevIsSCSIDevice(hostdev) &&
- hostdev->source.subsys.u.scsi.protocol !=
+ (hostdev->source.subsys.u.scsi.protocol !=
VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI));
}
--
2.20.1
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On Thu, Jan 10, 2019 at 06:40:33PM -0500, John Ferlan wrote: >It's essentially dead code. The only way hostdev->shareable >can be true is during virDomainHostdevDefParseXML when the >result of virHostdevIsSCSIDevice is true. While this might be true in our current codebase, I don't think this function should rely on the callers to only fill in hostdev->shareable for VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI to guard its usage of the hostdev->source union. Jano > >Signed-off-by: John Ferlan <jferlan@redhat.com> >--- > src/qemu/qemu_conf.c | 3 +-- > 1 file changed, 1 insertion(+), 2 deletions(-) > >diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c >index 20952e9607..945725566c 100644 >--- a/src/qemu/qemu_conf.c >+++ b/src/qemu/qemu_conf.c >@@ -1438,8 +1438,7 @@ static bool > qemuIsSharedHostdev(virDomainHostdevDefPtr hostdev) > { > return (hostdev->shareable && >- (virHostdevIsSCSIDevice(hostdev) && >- hostdev->source.subsys.u.scsi.protocol != >+ (hostdev->source.subsys.u.scsi.protocol != > VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI)); > } > >-- >2.20.1 > >-- >libvir-list mailing list >libvir-list@redhat.com >https://www.redhat.com/mailman/listinfo/libvir-list -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On 1/11/19 3:26 AM, Ján Tomko wrote: > On Thu, Jan 10, 2019 at 06:40:33PM -0500, John Ferlan wrote: >> It's essentially dead code. The only way hostdev->shareable >> can be true is during virDomainHostdevDefParseXML when the >> result of virHostdevIsSCSIDevice is true. > > While this might be true in our current codebase, > I don't think this function should rely on the callers > to only fill in hostdev->shareable for VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI > to guard its usage of the hostdev->source union. > > Jano > OK - I guess I was trying to be "compete" and partially considering that there's no reason this method wouldn't work for other hostdev types (PCI, USB, SCSI_HOST, MDEV) beyond of course the fact that for those types virDomainHostdevDefParseXML doesn't parse the readonly and shareable subelements. It's also not named qemuIsShareSCSIHostdev, but it does filter on SCSI. Still the first patch ends up being the fix and I separated them because I wasn't sure and figured I'd take another opinion on it! Should I "assume" patch1 is SFF... Tks - John >> >> Signed-off-by: John Ferlan <jferlan@redhat.com> >> --- >> src/qemu/qemu_conf.c | 3 +-- >> 1 file changed, 1 insertion(+), 2 deletions(-) >> >> diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c >> index 20952e9607..945725566c 100644 >> --- a/src/qemu/qemu_conf.c >> +++ b/src/qemu/qemu_conf.c >> @@ -1438,8 +1438,7 @@ static bool >> qemuIsSharedHostdev(virDomainHostdevDefPtr hostdev) >> { >> return (hostdev->shareable && >> - (virHostdevIsSCSIDevice(hostdev) && >> - hostdev->source.subsys.u.scsi.protocol != >> + (hostdev->source.subsys.u.scsi.protocol != >> VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI)); >> } >> >> -- >> 2.20.1 >> >> -- >> libvir-list mailing list >> libvir-list@redhat.com >> https://www.redhat.com/mailman/listinfo/libvir-list -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On Fri, Jan 11, 2019 at 07:44:38AM -0500, John Ferlan wrote: > > >On 1/11/19 3:26 AM, Ján Tomko wrote: >> On Thu, Jan 10, 2019 at 06:40:33PM -0500, John Ferlan wrote: >>> It's essentially dead code. The only way hostdev->shareable >>> can be true is during virDomainHostdevDefParseXML when the >>> result of virHostdevIsSCSIDevice is true. >> >> While this might be true in our current codebase, >> I don't think this function should rely on the callers >> to only fill in hostdev->shareable for VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI >> to guard its usage of the hostdev->source union. >> >> Jano >> > >OK - I guess I was trying to be "compete" and partially considering that >there's no reason this method wouldn't work for other hostdev types >(PCI, USB, SCSI_HOST, MDEV) beyond of course the fact that for those >types virDomainHostdevDefParseXML doesn't parse the readonly and >shareable subelements. > >It's also not named qemuIsShareSCSIHostdev, but it does filter on SCSI. >Still the first patch ends up being the fix and I separated them because >I wasn't sure and figured I'd take another opinion on it! > >Should I "assume" patch1 is SFF... > Yes, Jano -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
© 2016 - 2024 Red Hat, Inc.