Ján Tomko wrote: > On Mon, Nov 21, 2016 at 06:45:05PM +0300, Roman Bogorodskiy wrote: > >Use 'goto cleanup'-style error handling instead of explicitly > >freeing variables in every error path. > >--- > > src/bhyve/bhyve_command.c | 39 +++++++++++++++++---------------------- > > 1 file changed, 17 insertions(+), 22 deletions(-) > > > >diff --git a/src/bhyve/bhyve_command.c b/src/bhyve/bhyve_command.c > >index 022b03b..4914d98 100644 > >--- a/src/bhyve/bhyve_command.c > >+++ b/src/bhyve/bhyve_command.c > > >@@ -114,9 +104,14 @@ bhyveBuildNetArgStr(const virDomainDef *def, > > virCommandAddArgFormat(cmd, "%d:0,virtio-net,%s,mac=%s", > > net->info.addr.pci.slot, > > realifname, virMacAddrFormat(&net->mac, macaddr)); > >+ > >+ ret = 0; > >+ cleanup: > >+ VIR_FREE(brname); > > Okay, brname was leaked before. Yeah, I've noticed that while working on the e1000 support patches. > >+ VIR_FREE(net->ifname); > > But freeing stuff from the virDomainNetDef structure on success seems > wrong. > > Previously, all the error codepaths except > if (VIR_STRDUP(realifname, "tap0") < 0) > on a dry run left net->ifname at NULL, but after this patch it seems > the network will not be cleaned up later even though it was created > successfully. > > ACK if you wrap it in: > if (ret < 0) Pushed with this change squashed, thanks! Roman Bogorodskiy
Attachment:
signature.asc
Description: PGP signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list
