Re: [libvirt] [PATCH 3/4] Cache result of QEMU capabilities extraction

On 03/05/2014 10:53 AM, Daniel P. Berrange wrote: > Extracting capabilities from QEMU takes a notable amount of time > when all QEMU binaries are installed. Each system emulator > needs about 200-300ms multiplied by 26 binaries == ~5-8 seconds. > > This change causes the QEMU driver to save an XML file containing > the content of the virQEMUCaps object instance in the cache > dir eg /var/cache/libvirt/qemu/capabilities/$SHA256(binarypath).xml > or $HOME/.cache/libvirt/qemu/cache/capabilities/$SHA256(binarypath).xml > > We attempt to load this and only if it fails, do we fallback to > probing the QEMU binary. The timestamp of the file is compared to > the timestamp of the QEMU binary and discarded if the QEMU binary > is newer. > > Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com&gt; > --- > src/qemu/qemu_capabilities.c | 412 ++++++++++++++++++++++++++++++++++++++++++- > src/qemu/qemu_capabilities.h | 2 + > src/qemu/qemu_driver.c | 1 + > 3 files changed, 407 insertions(+), 8 deletions(-) > > @@ -44,6 +45,7 @@ > #include <unistd.h> > #include <sys/wait.h> > #include <stdarg.h> > +#include <utime.h> <utime.h> (and the utime() function) is deprecated by POSIX and therefore no longer portable because it corrupts sub-second timestamps. Better is to use gnulib's utimensat (or futimens). But what timestamps do we actually have to munge? > + > + ut.actime = qemuCaps->mtime; > + ut.modtime = qemuCaps->mtime; > + if (utime(filename, &ut) < 0) { > + virReportSystemError(errno, > + _("Failed to set mtime on '%s' for '%s'"), > + filename, qemuCaps->binary); Why not just store the qemu binary timestamp in the XML, rather than playing games with the mtime of the xml file?