On Mon, Sep 08, 2014 at 02:33:05PM +0100, Richard W.M. Jones wrote:
On Mon, Sep 08, 2014 at 07:46:34AM +0200, Martin Kletzander wrote:There were various problems introduced by the series on FD passing. The path for the socket was not created, the socket was not removed before binding it, etc. Let's see if it's any better this time. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=927369 Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1138604 Martin Kletzander (5): rpc: reformat the flow to make a bit more sense remove redundant pidfile path constructions util: fix potential leak in error codepath util: get rid of unnecessary umask() call rpc: make daemon spawning a bit more intelligent daemon/libvirtd.c | 41 ++------------ src/libvirt_private.syms | 1 + src/locking/lock_daemon.c | 42 ++------------- src/rpc/virnetsocket.c | 133 +++++++++++++++++++++++++++++++--------------- src/util/virpidfile.c | 48 ++++++++++++++++- src/util/virpidfile.h | 7 ++- 6 files changed, 151 insertions(+), 121 deletions(-)Patches 1 & 5 don't apply cleanly. Patch 5 is particularly quite different from current head.
Patch 1 doesn't apply cleanly because I've sent *only* the version formatted with '-w' (and not the usual one) and patch 5 depends on that, so that's probably the reason for it. I'll send a normal 1/5 so it can be applied cleanly, thank you for noticing. Martin
Attachment:
signature.asc
Description: Digital signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list