- Devel - libvirt List Archives

[Libvir] Re: libvirt on mingw
by Richard W.M. Jones 31 Jan '08

31 Jan '08

Brecht Sanders wrote: > I wasn't able to run virt-manager yet because of Python dependancy > issues, but I'm sure I'll be able to work that out. > Did you have virt-manager running on win32 yet? No, I wasn't able to get virt-manager running. Probably the same problems as you are having: building Python modules (specifically, libvirtmod.dll) which can work from the Win32 version of python, with gcc/MinGW, proves to be very hard. If you get it working, please let me know. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] virsh capabilities cmd: plug a leak
by Jim Meyering 31 Jan '08

31 Jan '08

Here's another: $ valgrind --leak-check=full \ ./virsh --connect test://$PWD/../docs/testnode.xml capabilities ... ==16392== 368 bytes in 1 blocks are definitely lost in loss record 13 of 27 ==16392== at 0x4A059F6: malloc (vg_replace_malloc.c:149) ==16392== by 0x387BC7CC11: strdup (strdup.c:43) ==16392== by 0x4194D2: testGetCapabilities (test.c:1006) ==16392== by 0x412B43: virConnectGetCapabilities (libvirt.c:2119) ==16392== by 0x408047: cmdCapabilities (virsh.c:2018) ==16392== by 0x40C5F9: vshCommandRun (virsh.c:4032) ==16392== by 0x40E6AF: main (virsh.c:4991) ==16392== Here's the patch: * src/virsh.c (cmdCapabilities): Plug a small leak. diff --git a/src/virsh.c b/src/virsh.c index c09dc8d..0d93fb6 100644 --- a/src/virsh.c +++ b/src/virsh.c @@ -2020,6 +2020,7 @@ cmdCapabilities (vshControl * ctl, vshCmd * cmd ATTRIBUTE_UNUSED) return FALSE; } vshPrint (ctl, "%s\n", caps); + free (caps); return TRUE; } -- 1.5.4.rc5.1.g0fa73

3 2

Re: [Libvir] Libvirt on windows
by Richard W.M. Jones 31 Jan '08

31 Jan '08

Gabriel Kaufmann wrote: > I tool libvirt-0.dll and I am using it in VC. > In order to do this, I had to dynamically load the DLL and map all its functions. This is somewhat outside my experience. My best suggestion (guess) is to try adding __declspec(dllexport) around some libvirt functions to see if this causes them to be exported to VC++-compiled programs. On gcc declspec doesn't appear to be necessary. See also: http://www.cygwin.com/ml/cygwin/2000-05/msg00430.html Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

2 2

[Libvir] Re: libvirt on mingw
by Richard W.M. Jones 30 Jan '08

30 Jan '08

Brecht Sanders wrote: > Hello again, > Note that I am using gcc (mingw) 3.4.5, and not 4 as you recommend. > The attached patch got rid of the warnings though. > Maybe there is a compiler switch to fix it in a cleaner way though. > Anyway, with the attached patch it does compile. > Basically it casts to void* before converting to a different type. Cheers for the patch, I'll take a look at it tomorrow ... There is a reason why you need to use gcc 4 although it's been so long since I did this that I can't remember exactly why now. Possibly it was some compiler feature that we use in libvirt. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

1 0

[Libvir] [PATCH] avoid virsh hang due to missing virDomainFree(dom) call
by Jim Meyering 30 Jan '08

30 Jan '08

Without the following patch, this command would hang printf 'domuuid fc4\ndomstate fc4\n' \ | ./virsh --connect test://$PWD/../docs/testnode.xml with this stack trace: __lll_lock_wait ... _L_lock_105 ... __pthread_mutex_lock ... virUnrefDomain (domain=0x6a8b30) at hash.c:884 virDomainFree (domain=0x6a8b30) at libvirt.c:1211 cmdDomstate (ctl=0x7fffea723390, cmd=0x6a8b10) at virsh.c:677 vshCommandRun (ctl=0x7fffea723390, cmd=0x6a8b10) at virsh.c:4033 main (argc=3, argv=0x7fffea7234e8) at virsh.c:5015 The problem is that cmdDomuuid didn't call virDomainFree(dom), so cmdDomstate hangs trying to do the Unref. I then did a mind-numbing audit of all of the other cmd* functions in virsh.c and found two other cases in which a "dom" may not be released. Here's the patch: Avoid virsh hang due to missing virDomainFree(dom) calls * src/virsh.c (cmdDomuuid): Add missing virDomainFree call. (cmdAttachDevice): Likewise. (cmdDetachDevice): Likewise. --- src/virsh.c | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/src/virsh.c b/src/virsh.c index 0d93fb6..487f256 100644 --- a/src/virsh.c +++ b/src/virsh.c @@ -2161,6 +2161,7 @@ cmdDomuuid(vshControl * ctl, vshCmd * cmd) else vshError(ctl, FALSE, "%s", _("failed to get domain UUID")); + virDomainFree(dom); return TRUE; } @@ -3038,8 +3039,10 @@ cmdAttachDevice(vshControl * ctl, vshCmd * cmd) return FALSE; } - if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) + if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) { + virDomainFree(dom); return FALSE; + } ret = virDomainAttachDevice(dom, buffer); free (buffer); @@ -3092,8 +3095,10 @@ cmdDetachDevice(vshControl * ctl, vshCmd * cmd) return FALSE; } - if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) + if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) { + virDomainFree(dom); return FALSE; + } ret = virDomainDetachDevice(dom, buffer); free (buffer); -- 1.5.4.rc5.1.g0fa73

2 5

[Libvir] plug small leaks upon failure
by Jim Meyering 30 Jan '08

30 Jan '08

with more than 20 domains, I get this: $ valgrind --leak-check=full ./virsh --connect test://$PWD/../docs/testnode.xml list ... error: failed to connect to the hypervisor ... ==13756== 5 bytes in 1 blocks are definitely lost in loss record 1 of 32 ==13756== at 0x4A059F6: malloc (vg_replace_malloc.c:149) ==13756== by 0x387BC7CC11: strdup (strdup.c:43) ==13756== by 0x41D9F2: virXPathString (xml.c:486) ==13756== by 0x4174D3: testLoadDomain (test.c:248) ==13756== by 0x417BA4: testLoadDomainFromFile (test.c:401) ==13756== by 0x418E57: testOpenFromFile (test.c:797) ==13756== by 0x4192CE: testOpen (test.c:913) ==13756== by 0x40FA20: do_open (libvirt.c:572) ==13756== by 0x40FDE4: virConnectOpenAuth (libvirt.c:681) ==13756== by 0x40D5B1: vshInit (virsh.c:4464) ==13756== by 0x40E67C: main (virsh.c:4985) ==13756== ==13756== ==13756== 12,996 (352 direct, 12,644 indirect) bytes in 1 blocks are definitely lost in loss record 31 of 32 ==13756== at 0x4A059F6: malloc (vg_replace_malloc.c:149) ==13756== by 0x38864880E6: xmlXPathNewContext (in /usr/lib64/libxml2.so.2.6.31) ==13756== by 0x417494: testLoadDomain (test.c:242) ==13756== by 0x417BA4: testLoadDomainFromFile (test.c:401) ==13756== by 0x418E57: testOpenFromFile (test.c:797) ==13756== by 0x4192CE: testOpen (test.c:913) ==13756== by 0x40FA20: do_open (libvirt.c:572) ==13756== by 0x40FDE4: virConnectOpenAuth (libvirt.c:681) ==13756== by 0x40D5B1: vshInit (virsh.c:4464) ==13756== by 0x40E67C: main (virsh.c:4985) ==13756== ==13756== LEAK SUMMARY: ==13756== definitely lost: 357 bytes in 2 blocks. ==13756== indirectly lost: 12,644 bytes in 32 blocks. ==13756== possibly lost: 0 bytes in 0 blocks. ==13756== still reachable: 20,614 bytes in 362 blocks. ==13756== suppressed: 0 bytes in 0 blocks. ==13756== Reachable blocks (those to which a pointer was found) are not shown. ==13756== To see them, rerun with: --leak-check=full --show-reachable Here's the fix: * test.c (testLoadDomain): Avoid leaks upon failure. --- src/test.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/test.c b/src/test.c index fe5da41..ab83f65 100644 --- a/src/test.c +++ b/src/test.c @@ -327,7 +327,7 @@ static int testLoadDomain(virConnectPtr conn, } } if (handle < 0) - return (-1); + goto error; privconn->domains[handle].active = 1; privconn->domains[handle].id = domid; -- 1.5.4.rc5.1.g0fa73

2 1

[Libvir] [PATCH] typo
by Atsushi SAKAI 30 Jan '08

30 Jan '08

Hi, I finally succeed for building libvirt on windows(MinGW/MSYS). Thank you for coopration. During that work I found several typo in libvirt document. Thanks Atsushi SAKAI

2 5

[Libvir] tiny patch, FYI: Let "make syntax-check" pass.
by Jim Meyering 30 Jan '08

30 Jan '08

FYI I've just checked this in: Let "make syntax-check" pass. * .x-sc_avoid_if_before_free: Exempt ChangeLog. * tests/statstest.c: Include <config.h>, not "config.h". diff --git a/.x-sc_avoid_if_before_free b/.x-sc_avoid_if_before_free index 627dcdf..f83ae7b 100644 --- a/.x-sc_avoid_if_before_free +++ b/.x-sc_avoid_if_before_free @@ -2,3 +2,4 @@ ^gnulib/lib/printf-parse\.c$ ^gnulib/lib/vasnprintf\.c$ ^build-aux/find-unnecessary-if-before-free$ +^ChangeLog$ diff --git a/tests/statstest.c b/tests/statstest.c index 7fa6f46..4c19833 100644 --- a/tests/statstest.c +++ b/tests/statstest.c @@ -1,4 +1,4 @@ -#include "config.h" +#include <config.h> #include <stdio.h> #include <stdlib.h>

1 0

Fwd: Re: [Libvir] update after changing XML
by Richard W.M. Jones 29 Jan '08

29 Jan '08

-- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903

3 3

[Libvir] VIR_DOMAIN_NOSTATE
by Russ Blaine 29 Jan '08

29 Jan '08

On Solaris-based xen 3.1.2 I am seeing virDomainGetInfo() often returning VIR_DOMAIN_NOSTATE in dominfo.state for domUs that are powered on and running. What exactly does VIR_DOMAIN_NOSTATE mean? Why does it exist? I suspect it is a catch-all in the API for hosts that don't entirely have their act together, that do not always return a valid guest state. Is that the case? Is anyone else seeing this on Xen? -- ----------------------------------------------------- Russ Blaine | Solaris Kernel | russell.blaine(a)sun.com

4 10