Re: [PATCH v6 10/10] tests: add support for LoongArch

[Xiaotian Wu]

在 2022/8/3 上午2:50, Glenn Washburn 写道:
> On Tue,  2 Aug 2022 15:13:37 +0800
> Xiaotian Wu <wuxiaotian@loongson.cn> wrote:
>
> > Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
> > ---
> >   tests/ahci_test.in       |  2 +-
> >   tests/ehci_test.in       |  2 +-
> >   tests/ohci_test.in       |  2 +-
> >   tests/pata_test.in       |  2 +-
> >   tests/uhci_test.in       |  2 +-
> >   tests/util/grub-shell.in | 14 ++++++++++++++
> >   6 files changed, 19 insertions(+), 5 deletions(-)
> >
> > diff --git a/tests/ahci_test.in b/tests/ahci_test.in
> > index 6d2e61d4e..70646a24e 100644
> > --- a/tests/ahci_test.in
> > +++ b/tests/ahci_test.in
> > @@ -30,7 +30,7 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" in
> >       mips*-arc | mips*-qemu_mips)
> >         exit 77;;
> >       # FIXME: No native drivers are available for those
> > -    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi)
> > +    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi | loongarch64-efi)
> >         exit 77;;
> >   esac
> >   
> > diff --git a/tests/ehci_test.in b/tests/ehci_test.in
> > index df671b4b6..bf823a5de 100644
> > --- a/tests/ehci_test.in
> > +++ b/tests/ehci_test.in
> > @@ -30,7 +30,7 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" in
> >       mips*-arc | mips*-qemu_mips)
> >         exit 77;;
> >       # FIXME: No native drivers are available for those
> > -    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi)
> > +    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi | loongarch64-efi)
> >         exit 77;;
> >   esac
> >   
> > diff --git a/tests/ohci_test.in b/tests/ohci_test.in
> > index 741ad881f..a40d3bc0a 100644
> > --- a/tests/ohci_test.in
> > +++ b/tests/ohci_test.in
> > @@ -30,7 +30,7 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" in
> >       mips*-arc | mips*-qemu_mips)
> >         exit 77;;
> >       # FIXME: No native drivers are available for those
> > -    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi)
> > +    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi | loongarch64-efi)
> >         exit 77;;
> >   esac
> >   
> > diff --git a/tests/pata_test.in b/tests/pata_test.in
> > index 31144a8fd..4d0e7d573 100644
> > --- a/tests/pata_test.in
> > +++ b/tests/pata_test.in
> > @@ -33,7 +33,7 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" in
> >       i386-efi)
> >         exit 77;;
> >       # FIXME: No native drivers are available for those
> > -    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi)
> > +    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi | loongarch64-efi)
> >         exit 77;;
> >       i386-ieee1275)
> >         disk=hdb
> > diff --git a/tests/uhci_test.in b/tests/uhci_test.in
> > index 5aa5eb726..de199a281 100644
> > --- a/tests/uhci_test.in
> > +++ b/tests/uhci_test.in
> > @@ -30,7 +30,7 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" in
> >       mips*-arc | mips*-qemu_mips)
> >         exit 77;;
> >       # FIXME: No native drivers are available for those
> > -    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi)
> > +    powerpc-ieee1275 | sparc64-ieee1275 | arm*-efi | loongarch64-efi)
> >         exit 77;;
> >   esac
> >   
> > diff --git a/tests/util/grub-shell.in b/tests/util/grub-shell.in
> > index 4828afb7c..e0cfe6f01 100644
> > --- a/tests/util/grub-shell.in
> > +++ b/tests/util/grub-shell.in
> > @@ -194,6 +194,16 @@ case "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" 
> > in
> >         disk="device virtio-blk-device,drive=hd1 -drive if=none,id=hd1,file="
> >         serial_port=efi0
> >         ;;
> > +    loongarch64-efi)
> > +       qemu=qemu-system-loongarch64
> > +       boot=hd
> > +       console=console
> > +       trim=1
> > +       qemuopts="-machine virt -cpu la464 -smp 2 -nographic -L ${srcdir} \
> > +                 -bios /usr/share/qemu/edk2-loongarch64-code.fd $qemuopts"
> The firmware should not be specified with an absolute path because that
> effectively disables searching for the firmware and make the -L option
> useless. So I believe it should just be "-bios edk2-loongarch64-code.fd".
Thanks,  you are right.
> > +       disk="device virtio-blk-pci,drive=hd1 -drive if=none,id=hd1,file="
> I realize now that I was confused by the hd1 here. This is just an identifier
> and has no other meaning. In fact, this disk gets set as the first disk
> in QEMU. Looking at the code again, the extra disk handling code really needs
> to be updated and reworked. I think this line is fine for now.
>
> > +       serial_port=
> > +       ;;
> >       *)
> >         boot=hd
> >         qemu=qemu-system-i386
> > @@ -395,6 +405,8 @@ fi
> >   if [ x$boot = xhd ]; then
> >       if [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = arm64-efi ] || [ 
> > "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = arm-efi ]; then
> >         device="device virtio-blk-device,drive=hd0 -drive if=none,id=hd0,file="
> > +    elif [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = 
> > loongarch64-efi ]; then
> > +       device="device virtio-blk-pci,drive=hd0 -drive if=none,id=hd0,file="
> The hd0 here is also misleading, as it will get set as the last of a series
> of disks, which in the case of the partmap_test will the the second disk.
> I know this is essentially copied from the if block above it, but instead
> of using hd0, how about using "grubdev", which seems a more accurate name.
>
> Also, I'm wondering why use "virtio-blk-pci" instead of "virtio-blk-device"
> as ARM does above? I this just the more preferred way to do this with QEMU
> now, or is there an important functional difference?

qemu-system-loongarch64 does not support 'virtio-blk-device'. If it is 
used, I will get an error:

qemu-system-loongarch64: -device virtio-blk-device,drive=grubdev: No 
'virtio-bus' bus found for device 'virtio-blk-device'

So, I can only use `virtio-blk-pci`.

> Glenn
>
> >       elif [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = mips-arc 
> > ]; then
> >         device="hdb "
> >       else
> > @@ -405,6 +417,8 @@ fi
> >   if [ x$boot = xcd ]; then
> >       if [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = arm64-efi ] || [ 
> > "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = arm-efi ]; then
> >         device="device virtio-blk-device,drive=cd0 -drive 
> > if=none,id=cd0,media=cdrom,file="
> > +    elif [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = 
> > loongarch64-efi ]; then
> > +       device="device virtio-blk-pci,drive=cd0 -drive 
> > if=none,id=cd0,media=cdrom,file="
> >       elif [ "${grub_modinfo_target_cpu}-${grub_modinfo_platform}" = powerpc-ieee1275 ] 
> > && [ x$pseries != xy ] ; then
> >         device="-drive if=ide,media=cdrom,file="
> >       else

--
Best Regards
Xiaotian Wu