Dan Smith [Thu, 26 Jun 2008 16:05:02 +0000 (16:05 +0000)]
[LXC] Detect support for NETNS in lxc driver initialization
Allow check for containers support to be done without CLONE_NEWNET, and then
determine support on the fly by checking for iproute2 support and a
successful clone(CLONE_NEWNET). This lets us set a flag for later, as well
as not completely disable LXC support on a system without NETNS support.
Daniel Veillard [Thu, 19 Jun 2008 14:39:49 +0000 (14:39 +0000)]
fix a couple of List functions for max = 0 issues
* src/xen_internal.c src/xend_internal.c: fix a couple of issues
for List functions where max given is 0
Daniel
Jim Meyering [Thu, 19 Jun 2008 14:30:52 +0000 (14:30 +0000)]
virsh fails to read files larger than BUFSIZ bytes
* src/util.c (fread_file_lim): Use VIR_REALLOC_N, not VIR_ALLOC_N.
Bug introduced in 895d0fdf5bef358fafb91c672609190b3088097b.
* tests/Makefile.am (test_scripts): Add read-bufsiz.
* tests/read-bufsiz: New test for the above.
Chris Lalancette [Thu, 19 Jun 2008 11:58:49 +0000 (11:58 +0000)]
For 0.4.3, danpb's new memory management scheme went into libvirt. This is
fine, except that is subtly alters the semantics of malloc(), calloc(), and
realloc(). In particular, if you say:
foo = malloc(0);
glibc will happily return a non-NULL pointer to you. However, with the new
memory management stuff, if you say:
foo = VIR_ALLOC(0);
you will actually get a NULL pointer back. Personally, I think this is a
dangerous deviation from malloc() semantics that everyone is used to, and is
indeed causing problems with the remote driver. The short of it is that the
remote driver allocates memory on behalf of the remote side using VIR_ALLOC_N,
and this call is returning NULL so that the NULL checks elsewhere in the code
fire and return failure.
The attached patch fixes this situation by removing the 0 checks from the memory
allocation paths, and just lets them fall through to the normal malloc(),
calloc(), or realloc() routines, restoring old semantics.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Thu, 19 Jun 2008 10:38:36 +0000 (10:38 +0000)]
When doing the conversion to danpb's new memory API, a small bug was
introduced into the qemudNetworkIfaceConnect() function. In particular, there
is a call:
if (VIR_ALLOC_N(vm->tapfds, vm->ntapfds+2) < 0)
goto no_memory;
However, the tapfds structure is used to track *all* of the tap fds, and is
called once for each network that is being attached to the domain. VIR_ALLOC_N
maps to calloc(). So the first network would work just fine, but if you had
more than one network, subsequent calls to this function would blow away the
stored fd's that were already there and fill them all in with zeros. This
causes multiple problems, from the qemu domains not starting properly to
improper cleanup on shutdown. The attached patch just changes the VIR_ALLOC_N()
to a VIR_REALLOC_N(), and everything is happy again.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Tue, 17 Jun 2008 12:49:37 +0000 (12:49 +0000)]
This rather large patch rewrites the virStorageBackendISCSIFindLUNs() function
to only rely on sysfs for finding LUNs, given a session number. Along the way,
it also fixes the bug where we wouldn't find LUNs for older kernels (with the
block:sda format), and also fixes a race condition where we could try
to find the LUN before udev has finished connecting it.
This patch fixes a few different bugs:
1) We weren't finding LUNs on pre 2.6.24 kernels
2) libvirtd sysfs scanning was racing with udev, so we didn't always see LUNs
3) We weren't showing the 0'th LUN, although that can be a valid LUN
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Tue, 17 Jun 2008 12:47:10 +0000 (12:47 +0000)]
A small bugfix; we only need to call the iscsiadm sendtarget command when we are
first logging in; we don't need to do it for logout. Move the sendtarget
command into the Login() function.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Tue, 17 Jun 2008 12:46:38 +0000 (12:46 +0000)]
Older versions of iscsiadm didn't support the "-P 0" flag to the "iscsiadm
--mode session" command. However, just running "iscsiadm --mode session" seems
to work on all version of iscsiadm commands back to FC-6, so just use that.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Tue, 17 Jun 2008 12:45:59 +0000 (12:45 +0000)]
In src/util.c, virLog is just a wrapper around fprintf(stderr). Make sure to
put line breaks at the end of lines that use virLog() (noticed during testing).
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Chris Lalancette [Tue, 17 Jun 2008 12:45:24 +0000 (12:45 +0000)]
This patch changes things around so that virStorageBackendRunProgRegex() does
*not* virStorageReportError() if the fork()/exec() process it spawned returned a
!= 0 exit code. Rather, it returns the exitcode in this case, and it is up to
the higher level to determine whether this is a fatal error or not. The use
case for this change is in the iSCSI stuff; older versions of iscsiadm tools
would return a failure when getting the session number, despite the command
succeeding.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Pass syntax check for XM block devices over to XenD
* src/xm_internal.c: Pass responsibility for checking syntax
of XM block devices over the XenD (see also Xen cset 17617).
(Hiroyuki Kaguchi).
virDomainMemoryPeek API
* include/libvirt/libvirt.h.in, src/libvirt.c, src/driver.h,
src/libvirt_sym.version: New virDomainMemoryPeek API.
* qemud/remote.c, qemud/remote_protocol.x, src/remote_internal.c:
Support for remote.
* src/qemu_driver.c: QEMU driver implementation of API.
* src/test.c: Test driver (null) implementation of API.
* docs/hvsupport.html.in: Document API.
* libvirt.spec.in: New path /var/cache/libvirt for temporary
storage of memory images.
* qemud/libvirtd.init.in: Remove any old temp files in
/var/cache/libvirt on restarts.
* src/Makefile.am: make install creates /var/cache/libvirt.
* configure.in: Detect mkdir -p.
Mon Jun 9 15:42:34 PST 2008 David L. Leskovec <dlesko@linux.vnet.ibm.com>
* src/lxc_driver.c: Make console element is output only. Always open new
PTY when starting a container.
Fix string overrun when storing console name in VM def struct
virDomainBlockPeek call
* configure.in: Document AC_SYS_LARGEFILE.
* docs/hvsupport.html.in: Document HV support for virDomainBlockPeek.
* include/libvirt/libvirt.h.in, src/driver.h, src/libvirt.c,
src/libvirt_sym.version: Add virDomainBlockPeek infrastructure.
* src/qemu_driver.c, src/test.c: Null versions of this call.
* src/xen_unified.c, src/xend_internal.c, src/xend_internal.h,
src/xm_internal.c, src/xm_internal.h: Xen implementation.
* tests/sexpr2xmldata/sexpr2xml-curmem.xml,
tests/sexpr2xmldata/sexpr2xml-no-source-cdrom.xml: XML output
has been reordered slightly in the Xen driver, but should be
functionally the same.
Fix a few issues related to restart of libvirtd with containers running.
Mon May 12 23:32:21 PST 2008 David L. Leskovec <dlesko@linux.vnet.ibm.com>
* src/lxc_driver.c: Add sanity of tty pid before kill()
Ignore ECHILD errors during VM cleanup
Call functions to store tty pid and cleanup tty pid file
* src/lxc_conf.h: Add function to verify container process exists
Add facilities to manage storing the tty forward process pid
* src/lxc_conf.c: Add function to verify container process exists
Call function to verify container process during config load
Add facilities to manage storing the tty forward process pid
Call function to load tty pid during load config
Deprecate the conn, dom and net fields in the _virError structure.
* include/libvirt/libvirt.h.in: Add VIR_DEPRECATED macro.
* src/internal.h: Internal code is allowed to use deprecated
features, so override VIR_DEPRECATED here.
* include/libvirt/virterror.h: Deprecate the conn, dom and net
fields in the _virError structure.
Fix type-punning warning in remote code.
* qemud/rpcgen_fix.pl: Case-insensitive string match on identifiers.
* qemud/remote_protocol.c: Update this generated file.
projects / libvirt.git / log