[Qemu-devel] [RFC] arm/virt: mark timer in fdt as armv8-compatible

[Qemu-devel] [RFC] arm/virt: mark timer in fdt as armv8-compatible

[hw.claudio]

From: Claudio Fontana <claudio.fontana@huawei.com>

Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
---
 hw/arm/virt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Is the timer armv8-compatible? I think it is, at least the virtual
timer seems to work as expected.

When looking up the timer information in the guest btw,
where the guest is running AArch64, does it make sense to look
for an armv8-timer compatible string?
Or is the "/timer" path standard enough that I should look for
the node using that fixed path instead?

Thanks,

Claudio

diff --git a/hw/arm/virt.c b/hw/arm/virt.c
index d6fffc7..8dbaf9d 100644
--- a/hw/arm/virt.c
+++ b/hw/arm/virt.c
@@ -245,7 +245,7 @@ static void fdt_add_timer_nodes(const VirtBoardInfo *vbi)
 
     qemu_fdt_add_subnode(vbi->fdt, "/timer");
     qemu_fdt_setprop_string(vbi->fdt, "/timer",
-                                "compatible", "arm,armv7-timer");
+                                "compatible", "arm,armv7-timer,armv8-timer");
     qemu_fdt_setprop_cells(vbi->fdt, "/timer", "interrupts",
                                GIC_FDT_IRQ_TYPE_PPI, 13, irqflags,
                                GIC_FDT_IRQ_TYPE_PPI, 14, irqflags,
-- 
1.8.5.3

Re: [Qemu-devel] [RFC] arm/virt: mark timer in fdt as armv8-compatible

[Claudio Fontana]

This is obviously wrong, we have to put a \0 to separate the two options.

But thee question remains, should we mark that as arm,armv8-timer compatible?

Thanks,

Claudio

On 15 September 2014 10:14,  <hw.claudio@gmail.com> wrote:
> From: Claudio Fontana <claudio.fontana@huawei.com>
>
> Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
> ---
>  hw/arm/virt.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> Is the timer armv8-compatible? I think it is, at least the virtual
> timer seems to work as expected.
>
> When looking up the timer information in the guest btw,
> where the guest is running AArch64, does it make sense to look
> for an armv8-timer compatible string?
> Or is the "/timer" path standard enough that I should look for
> the node using that fixed path instead?
>
> Thanks,
>
> Claudio
>
> diff --git a/hw/arm/virt.c b/hw/arm/virt.c
> index d6fffc7..8dbaf9d 100644
> --- a/hw/arm/virt.c
> +++ b/hw/arm/virt.c
> @@ -245,7 +245,7 @@ static void fdt_add_timer_nodes(const VirtBoardInfo *vbi)
>
>      qemu_fdt_add_subnode(vbi->fdt, "/timer");
>      qemu_fdt_setprop_string(vbi->fdt, "/timer",
> -                                "compatible", "arm,armv7-timer");
> +                                "compatible", "arm,armv7-timer,armv8-timer");
>      qemu_fdt_setprop_cells(vbi->fdt, "/timer", "interrupts",
>                                 GIC_FDT_IRQ_TYPE_PPI, 13, irqflags,
>                                 GIC_FDT_IRQ_TYPE_PPI, 14, irqflags,
> --
> 1.8.5.3
>

Re: [Qemu-devel] [RFC] arm/virt: mark timer in fdt as armv8-compatible

[Peter Maydell]

On 16 September 2014 10:31, Claudio Fontana <hw.claudio@gmail.com> wrote:
> This is obviously wrong, we have to put a \0 to separate the two options.
>
> But thee question remains, should we mark that as arm,armv8-timer compatible?

I think we should, but only if the CPU we're presenting to
the guest is a v8 CPU (ie the ARM_FEATURE_V8 bit is set for
the CPU).

-- PMM

Re: [Qemu-devel] [RFC] arm/virt: mark timer in fdt as armv8-compatible

[Claudio Fontana]

On 25 September 2014 19:15, Peter Maydell <peter.maydell@linaro.org> wrote:
> On 16 September 2014 10:31, Claudio Fontana <hw.claudio@gmail.com> wrote:
>> This is obviously wrong, we have to put a \0 to separate the two options.
>>
>> But thee question remains, should we mark that as arm,armv8-timer compatible?
>
> I think we should, but only if the CPU we're presenting to
> the guest is a v8 CPU (ie the ARM_FEATURE_V8 bit is set for
> the CPU).
>
> -- PMM

Ok I am sending a patch for this momentarily.

Claudio