Re: [Qemu-devel] [PATCH] hw/arm/boot: fix direct kernel boot setup

[Andrew Jones]

On Tue, Jun 18, 2019 at 12:02:37PM +0100, Peter Maydell wrote:
> On Tue, 18 Jun 2019 at 09:34, Andrew Jones <address@hidden> wrote:
> >
> > We need to check ram_end, not ram_size.
> >
> > Fixes: 852dc64d665f ("hw/arm/boot: Diagnose layouts that put initrd or
> > DTB off the end of RAM")
> > Signed-off-by: Andrew Jones <address@hidden>
> > ---
> >  hw/arm/boot.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/hw/arm/boot.c b/hw/arm/boot.c
> > index b2f93f6beff6..8a280ab3ed49 100644
> > --- a/hw/arm/boot.c
> > +++ b/hw/arm/boot.c
> > @@ -1109,7 +1109,7 @@ static void arm_setup_direct_kernel_boot(ARMCPU *cpu,
> >                               info->initrd_filename);
> >                  exit(1);
> >              }
> > -            if (info->initrd_start + initrd_size > info->ram_size) {
> > +            if (info->initrd_start + initrd_size > ram_end) {
> >                  error_report("could not load initrd '%s': "
> >                               "too big to fit into RAM after the kernel",
> >                               info->initrd_filename);
> > --
> > 2.20.1
> 
> Reviewed-by: Peter Maydell <address@hidden>
> 
> I think I missed this because my test case doesn't have an
> initrd -- direct kernel boot works fine if all you're
> passing to QEMU is the kernel... I think we could clarify
> the commit message a little:

I found it using kvm-unit-tests which uses a small initrd to
pass environment variables to the guest. All the tests started
to report FAIL.

> 
> hw/arm/boot: fix direct kernel boot with initrd
> 
> Fix the condition used to check whether the initrd fits
> into RAM; this meant we were spuriously refusing to do
> a direct boot of a kernel in some cases if an initrd
> was also passed on the command line.

Actually I think we need another fix for this error too. We
weren't actually refusing do direct boot the kernel, but we
should have been. We're missing the 'exit(1)' after the error
message.

> 
> ?
> 
> (if you agree I can just fix up the commit message when I apply it.)

I agree with the improved commit message if we also add the
'exit(1)', otherwise "refusing to boot" isn't the right thing
to say.

Thanks,
drew

> 
> thanks
> -- PMM
>