Matthias Bolte wrote: > 2010/1/18 Jim Meyering <jim@xxxxxxxxxxxx>: >> At first I was going to call virDomainDeviceDefFree only "if (dev)", >> but saw that it handles a NULL "dev" just fine, so it's better to skip >> the test altogether, just as we do for many other free-like functions. >> >> >From ea8511d709492f5cdc152a1eaccbccd05f036648 Mon Sep 17 00:00:00 2001 >> From: Jim Meyering <meyering@xxxxxxxxxx> >> Date: Mon, 18 Jan 2010 16:55:36 +0100 >> Subject: [PATCH] qemu_driver: don't leak a virDomainDeviceDef buffer >> >> * src/qemu/qemu_driver.c (qemudDomainAttachDevice): Don't leak "dev". ... > NACK. This will probably result in a segfault because you are freeing > memory that is still in use. > > Yes the toplevel dev leaks here, but for example > qemudDomainAttachNetDevice some lines above takes parts from the dev > struct an assigns them to other structs _without_ copying them. Thanks. That is nastily unintuitive and sounds a lot like a bug. I hope it's on the list of things to be fixed by Dan's patch. > I found this leak some time ago too, but gave up on fixing it as I > noticed how entangled this code is. -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list