Eduardo Habkost <[email protected]> writes: > Since commit 25f8dd9, qdev_prop_set_globals_for_type() stops > applying global properties on the first error. It is a leftover > from when QEMU exited on any error when applying global property. > Now we print a warning about the first error, bug ignore all > other global properties after it.
Actually, we've stopped since forever. Before commit 25f8dd9, that was okay. Since commit, it isn't. Suggest: qdev_prop_set_globals_for_type() stops applying global properties on the first error. It is a leftover from when QEMU exited on any error when applying global property. Commit 25f8dd9 changed the fatal error to a warning, but neglected to drop the stopping. Fix that. > For example, the following command-line will not set CPUID level > to 3, but will warn only about "x86_64-cpu.vendor" being ignored. > > $ ./x86_64-softmmu/qemu-system-x86_64 \ > -global x86_64-cpu.vendor=x \ > -global x86_64-cpu.level=3 > qemu-system-x86_64: Warning: global x86_64-cpu.vendor=x ignored: Property > '.vendor' doesn't take value 'x' > > Fix this by not returning from qdev_prop_set_globals_for_type() > on the first error. > > Cc: Markus Armbruster <[email protected]> > Reviewed-by: Markus Armbruster <[email protected]> > Reviewed-by: Marcel Apfelbaum <[email protected]> > Signed-off-by: Eduardo Habkost <[email protected]> > --- > hw/core/qdev-properties.c | 1 - > 1 file changed, 1 deletion(-) > > diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c > index e3b2184..c10edee 100644 > --- a/hw/core/qdev-properties.c > +++ b/hw/core/qdev-properties.c > @@ -1088,7 +1088,6 @@ static void qdev_prop_set_globals_for_type(DeviceState > *dev, > assert(prop->user_provided); > error_reportf_err(err, "Warning: global %s.%s=%s ignored: ", > prop->driver, prop->property, prop->value); > - return; > } > } > } Preferably with a clarified commit message: Reviewed-by: Markus Armbruster <[email protected]>
