Message ID | 1507564902-9000-2-git-send-email-ian.jackson@eu.citrix.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Ian Jackson writes ("[PATCH 1/8] xen: link against xentoolcore"): > From: Anthony PERARD <anthony.perard@citrix.com> > > Xen libraries 4.10 will include a new xentoolcore library, without > which xendevicemodel et al will not work. The xentoolcore library was just committed to xen.git#staging, so at least this patch (or something like it) should go into qemu RSN. Anthony, do you plan to cherry pick this into the xen-specific qemu branch used for osstest tests ? Thanks, Ian.
On Mon, Oct 09, 2017 at 05:28:08PM +0100, Ian Jackson wrote: > Ian Jackson writes ("[PATCH 1/8] xen: link against xentoolcore"): > > From: Anthony PERARD <anthony.perard@citrix.com> > > > > Xen libraries 4.10 will include a new xentoolcore library, without > > which xendevicemodel et al will not work. > > The xentoolcore library was just committed to xen.git#staging, so at > least this patch (or something like it) should go into qemu RSN. I don't think it is necessary to do anything in qemu. The linker should find on its own the new libxentoolcore as long as an option -Wl,-rpath-link= provide the right path to xentoolcore when building qemu from xen.git. In other cases, the pkg-config files should be enough (configure doesn't need to known about a new xentoolcore.pc file). > Anthony, do you plan to cherry pick this into the xen-specific qemu > branch used for osstest tests ? > > Thanks, > Ian.
Anthony PERARD writes ("Re: [PATCH 1/8] xen: link against xentoolcore"): > On Mon, Oct 09, 2017 at 05:28:08PM +0100, Ian Jackson wrote: > > The xentoolcore library was just committed to xen.git#staging, so at > > least this patch (or something like it) should go into qemu RSN. > > I don't think it is necessary to do anything in qemu. The linker should > find on its own the new libxentoolcore as long as an option > -Wl,-rpath-link= provide the right path to xentoolcore when building > qemu from xen.git. But, the -rpath-link= specification in libxendevicemodel refers not to the xen.git build tree, but to the eventual installed copy. In my tests, this does in fact break the build. > In other cases, the pkg-config files should be > enough (configure doesn't need to known about a new xentoolcore.pc > file). In that case, yes, the .pc works. Ian.
Anthony PERARD writes ("Re: [PATCH 1/8] xen: link against xentoolcore"): > I don't think it is necessary to do anything in qemu. The linker should > find on its own the new libxentoolcore as long as an option > -Wl,-rpath-link= provide the right path to xentoolcore when building > qemu from xen.git. In other cases, the pkg-config files should be > enough (configure doesn't need to known about a new xentoolcore.pc > file). It seems that the build still works, without this patch but without the rest of my series too, because we end up passing -L/u/iwj/work/xen.git/tools/../tools/libs/toolcore -Wl,-rpath-link=/u/iwj/work/xen.git/tools/../tools/libs/toolcore (or similar) on the qemu link line. So when ld links against libxenstore (say) it finds that xenstore needs toolcore and finds toolcore in the relevant paths. We still need this patch for the rest of the series, though. Ian.
On Thu, Oct 19, 2017 at 05:38:10PM +0100, Ian Jackson wrote: > Anthony PERARD writes ("Re: [PATCH 1/8] xen: link against xentoolcore"): > > I don't think it is necessary to do anything in qemu. The linker should > > find on its own the new libxentoolcore as long as an option > > -Wl,-rpath-link= provide the right path to xentoolcore when building > > qemu from xen.git. In other cases, the pkg-config files should be > > enough (configure doesn't need to known about a new xentoolcore.pc > > file). > > It seems that the build still works, without this patch but without > the rest of my series too, because we end up passing > -L/u/iwj/work/xen.git/tools/../tools/libs/toolcore > -Wl,-rpath-link=/u/iwj/work/xen.git/tools/../tools/libs/toolcore > (or similar) on the qemu link line. So when ld links against > libxenstore (say) it finds that xenstore needs toolcore and finds > toolcore in the relevant paths. > > We still need this patch for the rest of the series, though. Of course, I was only arguing that this patch on its own is not usefull. Do you need a signed-off-by or review-by from me? Since it looks like I'm the author.
Anthony PERARD writes ("Re: [PATCH 1/8] xen: link against xentoolcore"): > On Thu, Oct 19, 2017 at 05:38:10PM +0100, Ian Jackson wrote: > > We still need this patch for the rest of the series, though. > > Of course, I was only arguing that this patch on its own is not usefull. Right. I wanted to accurately reflect the situation in the commit message, hence the explanation. > Do you need a signed-off-by or review-by from me? Since it looks like > I'm the author. Indeed, so I don't think it makes sense for you to say you've reviewed it. Ian.
diff --git a/configure b/configure index fd7e3a5..6f691df 100755 --- a/configure +++ b/configure @@ -2072,7 +2072,7 @@ if test "$xen" != "no" ; then $($pkg_config --modversion xencontrol | sed 's/\./ /g') )" xen=yes xen_pc="xencontrol xenstore xenguest xenforeignmemory xengnttab" - xen_pc="$xen_pc xenevtchn xendevicemodel" + xen_pc="$xen_pc xenevtchn xendevicemodel xentoolcore" QEMU_CFLAGS="$QEMU_CFLAGS $($pkg_config --cflags $xen_pc)" libs_softmmu="$($pkg_config --libs $xen_pc) $libs_softmmu" LDFLAGS="$($pkg_config --libs $xen_pc) $LDFLAGS" @@ -2104,18 +2104,20 @@ EOF cat > $TMPC <<EOF && #undef XC_WANT_COMPAT_MAP_FOREIGN_API #include <xenforeignmemory.h> +#include <xentoolcore.h> int main(void) { xenforeignmemory_handle *xfmem; xfmem = xenforeignmemory_open(0, 0); xenforeignmemory_map2(xfmem, 0, 0, 0, 0, 0, 0, 0); + xentoolcore_restrict_all(0); return 0; } EOF - compile_prog "" "$xen_libs -lxendevicemodel $xen_stable_libs" + compile_prog "" "$xen_libs -lxendevicemodel $xen_stable_libs -lxentoolcore" then - xen_stable_libs="-lxendevicemodel $xen_stable_libs" + xen_stable_libs="-lxendevicemodel $xen_stable_libs -lxentoolcore" xen_ctrl_version=41000 xen=yes elif