search for: guestfs_disk_format

...file for qemu-img. */ + tmpfilefd = mkstemp (tmpfile); + if (tmpfilefd == -1) { + fprintf (stderr, "%s: mkstemp failed: %m\n", argv[0]); + exit (EXIT_FAILURE); + } + close (tmpfilefd); + g = guestfs_create (); /* Do something which forks qemu-img subprocess. */ - fmt = guestfs_disk_format (g, "/dev/null"); + fmt = guestfs_disk_format (g, tmpfile); + unlink (tmpfile); if (fmt == NULL) { /* Test failed. */ fprintf (stderr, "%s: unexpected failure of test, see earlier messages\n", -- 2.14.3

...s is a reasonable proxy for presence of the qemu-img -U option since both features were added within a few months. However the problem is that testing for this requires us to run qemu before launch is called. That has API issues because it's unclear what code like: g = guestfs_create (); guestfs_disk_format (g, "disk.img"); guestfs_set_hv (g, "my-weird-qemu"); guestfs_launch (g); should do (granted, it is a peculiar corner case and the caller probably gets what they deserve). > One possible idea can be to cache the qemu_data struct in the handle, > so the direct backend...

...oth features were added within a few months. > > > > However the problem is that testing for this requires us to run qemu > > before launch is called. That has API issues because it's unclear > > what code like: > > > > g = guestfs_create (); > > guestfs_disk_format (g, "disk.img"); > > guestfs_set_hv (g, "my-weird-qemu"); > > guestfs_launch (g); > > > > should do (granted, it is a peculiar corner case and the caller > > probably gets what they deserve). > > > > > One possible idea can be t...

...(guestfs_h *g, const char *filename, int64_t size, > > } > > } > > > > + /* With libvirt >= 6.0 the backing format must be specified. */ > > + if (backingfile != NULL && backingformat == NULL) { > > + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); > > + if (!backingformat) > > + return -1; > > + if (STREQ (backingformat, "unknown")) { > > + error (g, _("could not auto-detect the format of the backing file %s"), > > + backingfile); > > +...

...file); > + if (tmpfilefd == -1) { > + fprintf (stderr, "%s: mkstemp failed: %m\n", argv[0]); > + exit (EXIT_FAILURE); > + } > + close (tmpfilefd); > + > g = guestfs_create (); > > /* Do something which forks qemu-img subprocess. */ > - fmt = guestfs_disk_format (g, "/dev/null"); > + fmt = guestfs_disk_format (g, tmpfile); > + unlink (tmpfile); > if (fmt == NULL) { > /* Test failed. */ > fprintf (stderr, "%s: unexpected failure of test, see earlier messages\n", > -- > 2.14.3 > > ______________...

...nt clustersize = -1; @@ -302,6 +303,18 @@ disk_create_qcow2 (guestfs_h *g, const char *filename, int64_t size, } } + /* With libvirt >= 6.0 the backing format must be specified. */ + if (backingfile != NULL && backingformat == NULL) { + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); + if (!backingformat) + return -1; + if (STREQ (backingformat, "unknown")) { + error (g, _("could not auto-detect the format of the backing file %s"), + backingfile); + return -1; + } + } + /* Assemble the qemu-img comman...

...the qemu-img -U option > since both features were added within a few months. > > However the problem is that testing for this requires us to run qemu > before launch is called. That has API issues because it's unclear > what code like: > > g = guestfs_create (); > guestfs_disk_format (g, "disk.img"); > guestfs_set_hv (g, "my-weird-qemu"); > guestfs_launch (g); > > should do (granted, it is a peculiar corner case and the caller > probably gets what they deserve). > > > One possible idea can be to cache the qemu_data struct in the...

...+ * Caller must free the returned string. + */ +static char * +get_source_format_or_autodetect (guestfs_h *g, struct drive *drv) +{ + if (drv->src.format) + return safe_strdup (g, drv->src.format); + + if (drv->src.protocol == drive_protocol_file) { + char *format; + + format = guestfs_disk_format (g, drv->src.u.path); + if (!format) + return NULL; + + if (STREQ (format, "unknown")) { + error (g, _("could not auto-detect the format.\n" + "If the format is known, pass the format to libguestfs, eg. using the\n" +...

...nt change in libvirt it is no longer possible to have a backing file where libvirt will autodetect the format. See: https://bugzilla.redhat.com/show_bug.cgi?id=1798148#c3 Therefore the only way to actually do this now would be either for libguestfs to detect the format (there is an API for this: guestfs_disk_format), or else to hard code the format in the ./configure option. In any case I will need to revert this until we can think of something else (if you still need the feature) because it will simply break with libvirt >= 6.0. Rich. > Signed-off-by: Pavel Butsykin <pbutsykin@virtuozzo.com> &...

...> > } > > > > } > > > > > > > > + /* With libvirt >= 6.0 the backing format must be specified. */ > > > > + if (backingfile != NULL && backingformat == NULL) { > > > > + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); > > > > + if (!backingformat) > > > > + return -1; > > > > + if (STREQ (backingformat, "unknown")) { > > > > + error (g, _("could not auto-detect the format of the backing file %s"), > > >...

When the result guestfs_disk_format() is not known, free the string before returning an error. --- lib/launch-libvirt.c | 1 + 1 file changed, 1 insertion(+) diff --git a/lib/launch-libvirt.c b/lib/launch-libvirt.c index e4961f201..844023b80 100644 --- a/lib/launch-libvirt.c +++ b/lib/launch-libvirt.c @@ -197,6 +197,7 @@ get_source...

...+303,18 @@ disk_create_qcow2 (guestfs_h *g, const char *filename, int64_t size, > } > } > > + /* With libvirt >= 6.0 the backing format must be specified. */ > + if (backingfile != NULL && backingformat == NULL) { > + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); > + if (!backingformat) > + return -1; > + if (STREQ (backingformat, "unknown")) { > + error (g, _("could not auto-detect the format of the backing file %s"), > + backingfile); > + return -1; > + } > +...

...ename, int64_t size, > > > } > > > } > > > > > > + /* With libvirt >= 6.0 the backing format must be specified. */ > > > + if (backingfile != NULL && backingformat == NULL) { > > > + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); > > > + if (!backingformat) > > > + return -1; > > > + if (STREQ (backingformat, "unknown")) { > > > + error (g, _("could not auto-detect the format of the backing file %s"), > > > + backi...

...t; > > > } > > > > > > > > > > + /* With libvirt >= 6.0 the backing format must be specified. */ > > > > > + if (backingfile != NULL && backingformat == NULL) { > > > > > + backingformat = backingformat_free = guestfs_disk_format (g, backingfile); > > > > > + if (!backingformat) > > > > > + return -1; > > > > > + if (STREQ (backingformat, "unknown")) { > > > > > + error (g, _("could not auto-detect the format of the backing file %s&qu...

...t; + +int +main (int argc, char *argv[]) +{ + guestfs_h *g = guestfs_create (); + char *mem, *fmt; + + /* Make sure we're using > 1GB in the main process. */ + mem = calloc (2 * 1024, 1024 * 1024); + assert (mem != NULL); + + /* Do something which forks qemu-img subprocess. */ + fmt = guestfs_disk_format (g, "/dev/null"); + if (fmt == NULL) { + /* Test failed. */ + fprintf (stderr, "%s: unexpected failure of test, see earlier messages\n", + argv[0]); + exit (EXIT_FAILURE); + } + + if (STRNEQ (fmt, "raw")) { + /* Test failed. */ + fprintf (...

The existing mechanism was clunky, slow and used ~ 10 MB of local disk. Rich.

This is almost just refactoring, but I also set the memory limit to really 1 GB, and not 1×10⁹. --- src/info.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) diff --git a/src/info.c b/src/info.c index d7f45f0..616ef50 100644 --- a/src/info.c +++ b/src/info.c @@ -56,6 +56,7 @@ static yajl_val get_json_output (guestfs_h *g, const char *filename); static char

https://bugs.launchpad.net/qemu/+bug/1740364 --- lib/guestfs-internal.h | 3 +++ lib/handle.c | 2 ++ lib/info.c | 39 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 44 insertions(+) diff --git a/lib/guestfs-internal.h b/lib/guestfs-internal.h index adeb9478a..c66c55e70 100644 --- a/lib/guestfs-internal.h +++ b/lib/guestfs-internal.h @@ -510,6 +510,9 @@ struct

...* qemu.conf. There is no way to detect if this option is set, so we @@ -1129,15 +1170,14 @@ construct_libvirt_xml_disk (guestfs_h *g, * the users pass the format to libguestfs which will faithfully pass * that to libvirt and this function won't be used. */ - format = guestfs_disk_format (g, drv_priv->path); + format = guestfs_disk_format (g, drv_priv->u.path); if (!format) return -1; if (STREQ (format, "unknown")) { - error (g, _("could not auto-detect the format of '%s'\n" + error (g, _("could not auto-detect...

QEMU upstream has broken snapshot=on ... again. These two patches stop using it entirely. Instead we run 'qemu-img create' to create overlay disks as required. Note that the libvirt and UML backends were already doing this: The libvirt backend because <transient/> has never worked, and the UML backend was running uml_mkcow because the UML-equivalent syntax of snapshot=on was