On Tue, 7 Nov 2023 at 18:18, BALATON Zoltan <bala...@eik.bme.hu> wrote: > > On Tue, 7 Nov 2023, Peter Maydell wrote: > > On Tue, 7 Nov 2023 at 18:05, BALATON Zoltan <bala...@eik.bme.hu> wrote: > >> So if it's tests/qtest/test-hmp.c that seems to try to run a bunch of > >> command for each machine AFAIU. This machine needs a firmware image but > >> this test seems to run it with -S and never starts the machine so could > >> take any file with -bios for that, e.g. -bios vof.bin (thought about -bios > >> /dev/null but that does not work but any other file should probably be > >> OK). I don't know how to tell test-hmp to use that thogh or to skip this > >> machine. > >> > >> At some point pegasos2 also needed a firmware but this wasn't a problem > >> back then. Either this test is new or we could just ignore this error? > > > > The test has been around for a long time, and no, you can't ignore > > 'make check' failures, they will break the CI. It's one of > > the tests we have that test properties that must be true for > > all board types (or all devices, in the case of some other > > tests). One of the properties is "it ought to start up cleanly > > without extra arguments". > > I'll send a patch in a minute which hopefully fixes that skipping the exit > when using qtest and called without -bios.
Yeah, looking at eg hw/mips/malta.c that's what we do there: we check qtest_enabled() before printing the error about not being able to load a BIOS blob. thanks -- PMM
