| To: | David Daney <ddaney@caviumnetworks.com> |
|---|---|
| Subject: | Re: [PATCH] serial: Add Cavium OCTEON UART definitions. |
| From: | Arnd Bergmann <arnd@arndb.de> |
| Date: | Tue, 7 Oct 2008 11:34:50 +0200 |
| Cc: | linux-serial@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mips@linux-mips.org, "Paoletti, Tomaso" <Tomaso.Paoletti@caviumnetworks.com> |
| In-reply-to: | <48EAAF97.8050307@caviumnetworks.com> |
| Original-recipient: | rfc822;linux-mips@linux-mips.org |
| References: | <48EAAF97.8050307@caviumnetworks.com> |
| Sender: | linux-mips-bounce@linux-mips.org |
| User-agent: | KMail/1.9.9 |
On Tuesday 07 October 2008, David Daney wrote:
>
> - up->port.type = PORT_16550A;
> +#ifdef CONFIG_CPU_CAVIUM_OCTEON
> + /* UPF_FIXED_PORT indicates an internal UART. */
> + if (up->port.flags & UPF_FIXED_PORT)
> + up->port.type = PORT_OCTEON;
> + else
> +#endif
> + up->port.type = PORT_16550A;
> +
This looks somewhat wrong, IMHO a device driver should not assume that
a CONFIG_CPU_* symbol is exclusive. You could have (maybe not now, but
in the future) a kernel that supports running on an Octeon as well
as some other Mips64 processor, and have UPF_FIXED_PORT uart on some
other machine, which will make the kernel think it is a PORT_OCTEON.
Arnd <><
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | [PATCH] serial: Add Cavium OCTEON UART definitions., David Daney |
|---|---|
| Next by Date: | Re: [PATCH] serial: Add Cavium OCTEON UART definitions., Alan Cox |
| Previous by Thread: | [PATCH] serial: Add Cavium OCTEON UART definitions., David Daney |
| Next by Thread: | Re: [PATCH] serial: Add Cavium OCTEON UART definitions., Alan Cox |
| Indexes: | [Date] [Thread] [Top] [All Lists] |