Re: [PATCH 0/3] sparc: port to copy_thread_tls() and struct kernel_clone_args

[Al Viro]

On Sun, May 17, 2020 at 05:34:34PM +0100, Mark Cave-Ayland wrote:

> FWIW if you're running a more recent version of QEMU (>=3.1) then you can also boot
> from the virtio-blk-pci device directly instead of having to switch back to the IDE
> device after installation as you have done above. Should be something like:
> 
> qemu-system-sparc64 \
>          -m 4096 \
>    	-device virtio-blk-pci,bus=pciB,drive=hd \
>  	-drive
> file=/home/brauner/.local/share/qemu/sparc64.img,format=raw,if=none,id=hd,bootindex=0 \
>         -net nic \
>  	-net user \
>  	-nographic
> 
> Note the removal of the legacy -boot argument and the addition of "bootindex=0" to
> the -drive argument.

	Is virtio-blk-pci more resilent to lost interrupt bug introduced in 
"sun4u: update PCI topology to include simba PCI bridges"?  I hadn't tried
it yet (reverted to the last working mainline qemu commit for now); IDE
definitely is screwed by that - both the Linux and NetBSD drivers, actually.

	A 50Mb worth of wget(1) is more than enough to trigger that crap;

commit 063833a6ec
Merge: d634fc0499 bcf9e2c2f2
Author: Peter Maydell <peter.maydell@xxxxxxxxxx>
Date:   Thu Oct 19 18:42:51 2017 +0100

    Merge remote-tracking branch 'remotes/mcayland/tags/qemu-sparc-signed' into staging

hangs, d634fc0499 works, bcf9e2c2f2 hangs.

I hadn't looked into details (the branch itself is only two commits long, but it
incorporates an openbios update - 35 commits there, some obviously pci- and
sun4u-related), but it's really easy to reproduce - -m 1024 and -hda <image>
are probably the only relevant arguments.  Even dd if=/dev/sda of=/dev/null bs=64m
is often enough to hang it, so I rather doubt that networking (e1000 on pciB,
FWIW, with tap for backend) has anything to do with that.