Re: [Qemu-devel] [PATCH 2/3] hw/arm/virt: Add another UART to the virt board

From: Peter Maydell <peter.maydell@linaro.org>
To: Andrew Jones <drjones@redhat.com>
Cc: qemu-arm <qemu-arm@nongnu.org>,
	QEMU Developers <qemu-devel@nongnu.org>,
	"Jason A . Donenfeld" <Jason@zx2c4.com>,
	Shannon Zhao <shannon.zhao@linaro.org>
Subject: Re: [Qemu-devel] [PATCH 2/3] hw/arm/virt: Add another UART to the virt board
Date: Tue, 12 Dec 2017 14:50:50 +0000	[thread overview]
Message-ID: <CAFEAcA-DyuObZjx-g2k9SkofLZ_akadfGYZCMum=3hQQLEdN_w@mail.gmail.com> (raw)
In-Reply-To: <20171212144513.sutjfenbjh33fm5n@kamzik.brq.redhat.com>

On 12 December 2017 at 14:45, Andrew Jones <drjones@redhat.com> wrote:
> Should this be
>
>     if (!vmc->no_second_uart && serial_hds[uart_count] != NULL) {
>
> A bit late now, but maybe the other UART resource allocations should have
> been conditional on actually being wired up too?

What does x86 do?

I think the conclusion we came to about UARTs in general was that
boards should always create the devices, and the UART models
should cope with being passed a NULL chardev backend. (This
is like real hardware, where the 2nd UART always exists even
if you haven't plugged a cable into the 2nd serial port on the
back of the box.)

thanks
-- PMM