November 2014 - Libguestfs - libguestfs List Archives

Re: [Libguestfs] Add support for slackware in supermin
by Richard W.M. Jones 15 Nov '14

15 Nov '14

On Sat, Nov 08, 2014 at 02:27:49PM +0100, Mathieu Bouillaguet wrote: > Hi, > > What needs to be implemented to add support for slackware distribution in > supermin ? hi Mathieu I'm on holiday at the moment and will give a more complete reply when I get back. In brief you need to be able to do a few simple package database meta operations such as listing all files in a package and listing the installed deps of a named pkg. This pre- supposes that slackware has such a packaging db. Pino CCd can give more details, and you can look at existing pkg systems in src/*.ml. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top

3 12

[PATCH 0/4 v3] readline escaping functions
by Maros Zatko 13 Nov '14

13 Nov '14

Helper functions for future support of backslash escaped spaces in filenames. There are a few tests too. Changed according to review remarks and fixed few other mistakes. Maros Zatko (4): fish: copy parse_quoted_string and hexdigit from fish.h to rl.c fish: rl.{c,h} - escaping functions for readline fish: basic tests for readline escaping autotools: add fish/test Makefile.am | 1 + configure.ac | 1 + fish/rl.c | 157 ++++++++++++++++++++++++++++++++++++++++++++++++ fish/rl.h | 25 ++++++++ fish/test/Makefile.am | 34 +++++++++++ fish/test/testquoting.c | 96 +++++++++++++++++++++++++++++ 6 files changed, 314 insertions(+) create mode 100644 fish/rl.c create mode 100644 fish/rl.h create mode 100644 fish/test/Makefile.am create mode 100644 fish/test/testquoting.c -- 1.9.3

1 4

libguestfs-tools with libvirt SASL authentication
by Dan Ryder (daryder) 13 Nov '14

13 Nov '14

Hello, I'm running into an issue using the 'virt-df' command when SASL is enabled in libvirt. I'm running version 1.26.5 of libguestfs on an Ubuntu 14.04 OS. I'm running 'virt-df' for all guests at one time, not specifying a domain or image to use. I was expecting a SASL authentication prompt but to no avail. Below is a sample run: root@all-in-one:~# virt-df -P 15 libvirt: XML-RPC error : authentication failed: Failed to start SASL negotiation: -4 (SASL(-4): no mechanism available: No worthy mechs found) virt-df: could not connect to libvirt (code 45, domain 7): authentication failed: Failed to start SASL negotiation: -4 (SASL(-4): no mechanism available: No worthy mechs found) However, if I do specify a domain to run on, I am prompted for SASL authentication: root@all-in-one:~# virt-df -d instance-0000004d libvirt needs authentication to connect to libvirt URI NULL (see also: http://libvirt.org/auth.html http://libvirt.org/uri.html) Please enter your authentication name: nova@all-in-one Please enter your password: Filesystem 1K-blocks Used Available Use% instance-0000004d:/dev/sdb 458 458 0 100% instance-0000004d:/dev/sda1 22773 12223 9322 54% Is there any way to successfully run 'virt-df' for all instances at once while SASL is enabled in libvirt? Thanks, Dan Ryder

2 4

[PATCH] daemon: check xfs label lengths (RHBZ#1162966).
by Pino Toscano 13 Nov '14

13 Nov '14

Similar to commit 52f9cd4882135910ea06e1e50ac6441d455c9ab1, but for xfs filesystems. --- daemon/daemon.h | 4 ++++ daemon/labels.c | 6 ++++++ daemon/xfs.c | 6 ++++++ 3 files changed, 16 insertions(+) diff --git a/daemon/daemon.h b/daemon/daemon.h index 0ccbc9e..f442efd 100644 --- a/daemon/daemon.h +++ b/daemon/daemon.h @@ -253,6 +253,10 @@ extern void main_loop (int sock) __attribute__((noreturn)); /*-- in xattr.c --*/ extern int copy_xattrs (const char *src, const char *dest); +/*-- in xfs.c --*/ +/* Documented in xfs_admin(8). */ +#define XFS_LABEL_MAX 12 + /* ordinary daemon functions use these to indicate errors * NB: you don't need to prefix the string with the current command, * it is added automatically by the client-side RPC stubs. diff --git a/daemon/labels.c b/daemon/labels.c index f417c57..cfcb4df 100644 --- a/daemon/labels.c +++ b/daemon/labels.c @@ -118,6 +118,12 @@ xfslabel (const char *device, const char *label) return -1; } + if (strlen (label) > XFS_LABEL_MAX) { + reply_with_error ("%s: xfs labels are limited to %d bytes", + label, XFS_LABEL_MAX); + return -1; + } + r = command (NULL, &err, str_xfs_admin, "-L", label, device, NULL); if (r == -1) { reply_with_error ("%s", err); diff --git a/daemon/xfs.c b/daemon/xfs.c index 725f7b3..687013b 100644 --- a/daemon/xfs.c +++ b/daemon/xfs.c @@ -498,6 +498,12 @@ do_xfs_admin (const char *device, } if (optargs_bitmask & GUESTFS_XFS_ADMIN_LABEL_BITMASK) { + if (strlen (label) > XFS_LABEL_MAX) { + reply_with_error ("%s: xfs labels are limited to %d bytes", + label, XFS_LABEL_MAX); + return -1; + } + ADD_ARG (argv, i, "-L"); ADD_ARG (argv, i, label); } -- 1.9.3

2 1

[PATCH] configure: simplify the qemu cpu mapping
by Pino Toscano 13 Nov '14

13 Nov '14

Use AS_CASE to properly map from the host cpu to the qemu cpu suffix; this avoids doing multiple string replacements using sed, and it is possible to use case-like syntax. --- configure.ac | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) diff --git a/configure.ac b/configure.ac index a571de5..75dfb37 100644 --- a/configure.ac +++ b/configure.ac @@ -755,15 +755,13 @@ AC_DEFINE_UNQUOTED([UNZIP],["$UNZIP"],[Name of unzip program.]) dnl Check for QEMU for running binaries on this $host_cpu, fall dnl back to basic 'qemu'. Allow the user to override it. -qemu_system="$( - echo qemu-system-$host_cpu | - $SED -e 's/-i@<:@456@:>@86/-i386/g' \ - -e 's/-arm.*/-arm/g' \ - -e 's/-amd64/-x86_64/g' \ - -e 's/-powerpc64/-ppc64/g' \ - -e 's/-ppc64le/-ppc64/g' - )" -default_qemu="qemu-kvm kvm $qemu_system qemu" +AS_CASE([$host_cpu], + [i@<:@456@:>@86],[qemu_cpu=i386], + [arm*],[qemu_cpu=arm], + [amd64],[qemu_cpu=x86_64], + [powerpc64 | ppc64le],[qemu_cpu=ppc64], + [qemu_cpu=$host_cpu]) +default_qemu="qemu-kvm kvm qemu-system-$qemu_cpu qemu" AC_ARG_WITH([qemu], [AS_HELP_STRING([--with-qemu="bin1 bin2 ..."], [set default QEMU binary @<:@default="[qemu-kvm] qemu-system-<host> qemu"@:>@])], -- 1.9.3

2 1

[PATCH] use pkg-config to look up ncurses
by Mike Frysinger 13 Nov '14

13 Nov '14

Hardcoding -ltinfo breaks on distros that do not build the sep library (which is the default ncurses behavior). Use pkg-config to look up the right libraries regardless of how the distro built things. --- builder/Makefile.am | 5 +++-- configure.ac | 18 ++++++------------ customize/Makefile.am | 2 +- mllib/Makefile.am | 6 +++--- resize/Makefile.am | 2 +- sparsify/link.sh.in | 2 +- sysprep/Makefile.am | 2 +- v2v/link.sh.in | 2 +- 8 files changed, 17 insertions(+), 22 deletions(-) diff --git a/builder/Makefile.am b/builder/Makefile.am index 7b66afb..e64a696 100644 --- a/builder/Makefile.am +++ b/builder/Makefile.am @@ -24,7 +24,8 @@ AM_CFLAGS = \ -I$(top_srcdir)/src \ -I$(top_srcdir)/fish \ -pthread \ - $(LIBLZMA_CFLAGS) + $(LIBLZMA_CFLAGS) \ + $(LIBCURSES_CFLAGS) EXTRA_DIST = \ $(SOURCES) \ @@ -172,7 +173,7 @@ OCAMLCLIBS = \ -L../src/.libs -lutils \ -L../gnulib/lib/.libs -lgnu \ -pthread -lpthread \ - -ltinfo -lcrypt + $(LIBCURSES_LIBS) -lcrypt if HAVE_OCAMLOPT virt-builder: $(OBJECTS) diff --git a/configure.ac b/configure.ac index a571de5..1c38084 100644 --- a/configure.ac +++ b/configure.ac @@ -352,19 +352,13 @@ struct sockaddr_un myaddr; AC_DEFINE_UNQUOTED([UNIX_PATH_MAX], $unix_path_max, [Custom value for UNIX_PATH_MAX]) ]) -dnl tgetent, tputs and UP [sic] are all required. They come from libtinfo -dnl which is pulled in as a dependency of libncurses. -old_LIBS="$LIBS" -AC_CHECK_LIB([tinfo], [tgetent], [], [ - AC_MSG_ERROR(['libtinfo' library (usually part of ncurses) not found, or it doesn't contain 'tgetent']) -]) -AC_CHECK_LIB([tinfo], [tputs], [], [ - AC_MSG_ERROR(['libtinfo' library (usually part of ncurses) not found, or it doesn't contain 'tputs']) +dnl tgetent, tputs and UP [sic] are all required. They come from the lower +dnl tinfo library, but might be part of ncurses directly. +PKG_CHECK_MODULES([LIBCURSES], [tinfo], [], [ + PKG_CHECK_MODULES([LIBCURSES], [ncurses]) ]) -AC_CHECK_LIB([tinfo], [UP], [], [ - AC_MSG_ERROR(['libtinfo' library (usually part of ncurses) not found, or it doesn't contain 'UP']) -]) -LIBS="$old_LIBS" +AC_SUBST([LIBCURSES_CFLAGS]) +AC_SUBST([LIBCURSES_LIBS]) dnl GNU gettext tools (optional). AC_CHECK_PROG([XGETTEXT],[xgettext],[xgettext],[no]) diff --git a/customize/Makefile.am b/customize/Makefile.am index f9dccd5..64509fd 100644 --- a/customize/Makefile.am +++ b/customize/Makefile.am @@ -124,7 +124,7 @@ OCAMLCFLAGS = $(OCAML_FLAGS) $(OCAML_WARN_ERROR) $(OCAMLPACKAGES) OCAMLOPTFLAGS = $(OCAMLCFLAGS) OCAMLCLIBS = \ - $(LIBXML2_LIBS) $(LIBINTL) -ltinfo -lcrypt \ + $(LIBXML2_LIBS) $(LIBINTL) $(LIBCURSES_LIBS) -lcrypt \ -L../src/.libs -lutils \ -L../gnulib/lib/.libs -lgnu diff --git a/mllib/Makefile.am b/mllib/Makefile.am index d248a72..112b610 100644 --- a/mllib/Makefile.am +++ b/mllib/Makefile.am @@ -99,7 +99,7 @@ OCAMLCFLAGS = $(OCAML_FLAGS) $(OCAML_WARN_ERROR) $(OCAMLPACKAGES) OCAMLOPTFLAGS = $(OCAMLCFLAGS) OCAMLCLIBS = \ - $(LIBXML2_LIBS) $(LIBINTL) -ltinfo -lcrypt \ + $(LIBXML2_LIBS) $(LIBINTL) $(LIBCURSES_LIBS) -lcrypt \ -L../src/.libs -lutils \ -L../gnulib/lib/.libs -lgnu @@ -154,11 +154,11 @@ check_SCRIPTS = common_utils_tests if HAVE_OCAMLOPT common_utils_tests: common_gettext.cmx tty-c.o tTY.cmx common_utils.cmx common_utils_tests.cmx $(OCAMLFIND) ocamlopt $(OCAMLOPTFLAGS) \ - mlguestfs.cmxa -linkpkg $^ -cclib -ltinfo -o $@ + mlguestfs.cmxa -linkpkg $^ -cclib $(LIBCURSES_LIBS) -o $@ else common_utils_tests: common_gettext.cmo tty-c.o tTY.cmo common_utils.cmo common_utils_tests.cmo $(OCAMLFIND) ocamlc $(OCAMLCFLAGS) \ - mlguestfs.cma -linkpkg $^ -cclib -ltinfo -custom -o $@ + mlguestfs.cma -linkpkg $^ -cclib $(LIBCURSES_LIBS) -custom -o $@ endif TESTS_ENVIRONMENT = $(top_builddir)/run --test diff --git a/resize/Makefile.am b/resize/Makefile.am index f7aa9b4..3030366 100644 --- a/resize/Makefile.am +++ b/resize/Makefile.am @@ -71,7 +71,7 @@ OCAMLCFLAGS = $(OCAML_FLAGS) $(OCAML_WARN_ERROR) $(OCAMLPACKAGES) OCAMLOPTFLAGS = $(OCAMLCFLAGS) OCAMLCLIBS = \ - $(LIBXML2_LIBS) $(LIBINTL) -ltinfo \ + $(LIBXML2_LIBS) $(LIBINTL) $(LIBCURSES_LIBS) \ -L../src/.libs -lutils \ -L../gnulib/lib/.libs -lgnu diff --git a/sparsify/link.sh.in b/sparsify/link.sh.in index 537e967..5d69237 100644 --- a/sparsify/link.sh.in +++ b/sparsify/link.sh.in @@ -19,4 +19,4 @@ # Hack automake to link binary properly. There is no other way to add # the -cclib parameter to the end of the command line. -exec "$@" -linkpkg -cclib '-ltinfo' +exec "$@" -linkpkg -cclib '@LIBCURSES_LIBS@' diff --git a/sysprep/Makefile.am b/sysprep/Makefile.am index 70f677d..086f14c 100644 --- a/sysprep/Makefile.am +++ b/sysprep/Makefile.am @@ -136,7 +136,7 @@ OCAMLCFLAGS = $(OCAML_FLAGS) $(OCAML_WARN_ERROR) $(OCAMLPACKAGES) OCAMLOPTFLAGS = $(OCAMLCFLAGS) OCAMLCLIBS = \ - $(LIBXML2_LIBS) $(LIBINTL) -ltinfo -lcrypt \ + $(LIBXML2_LIBS) $(LIBINTL) $(LIBCURSES_LIBS) -lcrypt \ -L../src/.libs -lutils \ -L../gnulib/lib/.libs -lgnu diff --git a/v2v/link.sh.in b/v2v/link.sh.in index d95a1dc..7ac7ee6 100644 --- a/v2v/link.sh.in +++ b/v2v/link.sh.in @@ -19,4 +19,4 @@ # Hack automake to link binary properly. There is no other way to add # the -cclib parameter to the end of the command line. -exec "$@" -linkpkg -cclib '-lutils -ltinfo -lcrypt @LIBVIRT_LIBS@ @LIBXML2_LIBS@ @LIBINTL@ -lgnu' +exec "$@" -linkpkg -cclib '-lutils @LIBCURSES_LIBS@ -lcrypt @LIBVIRT_LIBS@ @LIBXML2_LIBS@ @LIBINTL@ -lgnu' -- 2.1.3

2 1

[libhivex] Memory leak in hivex_node_delete_child?
by Mahmoud Al-Qudsi 13 Nov '14

13 Nov '14

Hello, I’m not able to pin it myself from a brief look, but it seems that under certain conditions a call to hivex_node_delete_child can cause the allocations from _hivex_get_children to not be freed properly? I know that if the return value of _hivex_get_children is -1, no free is called; but from what I can see, -1 means there is nothing to be freed/it’s been taken care of. I’ve attached the valgrind log together with the output of HIVEX_OPEN_DEBUG, please let me know if there’s anything else I can provide to make this easier. Many thanks, Mahmoud Al-Qudsi NeoSmart Technologies

3 3

Re: [Libguestfs] [libhivex] Undefined behavior when accessing invalid (too small) registry hives
by Mahmoud Al-Qudsi 13 Nov '14

13 Nov '14

> On Nov 11, 2014, at 1:57 AM, Richard W.M. Jones <rjones(a)redhat.com> wrote: > > Yes I was also meaning to do that after reading lcamtuf's postings. Yup. That's the one. > I just started a run now .. Will let it run for a few days and report > any issues on the list. Thank you. Do you mind running it under valgrind to catch out-of-bound reads? Mahmoud

2 1

Re: [Libguestfs] [nbdkit] leak? (#1)
by Richard W.M. Jones 13 Nov '14

13 Nov '14

On Thu, May 29, 2014 at 10:39:49AM -0700, knokknok wrote: > Not clear to me if the following malloc is freed within lzma or if this is a leak... > https://github.com/libguestfs/nbdkit/blob/master/plugins/xz/xzfile.c#L454 Sorry for the late reply - we don't use github for problem tracking so no one was looking at this. In this function (xzfile_read_block) the 'data' pointer is returned to the caller, who must free it. The calling code is a bit complicated because it puts the 'data' pointer into a cache, where it will be freed when the cache is deleted or the block in the cache expires. So I believe this is safe, but if you think otherwise, then let us know on the mailing list (libguestfs(a)redhat.com) Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top

1 0

Re: [Libguestfs] [libhivex] Memory leak in hivex_node_delete_child?
by Mahmoud Al-Qudsi 11 Nov '14

11 Nov '14

> On Nov 11, 2014, at 5:36 AM, Mahmoud Al-Qudsi <mqudsi(a)neosmart.net> wrote: > I'll test it and be back. And it works perfectly. Thanks, Mahmoud

1 0