linux-kernel - Re: [Patch v4] ARC: Dynamically determine BASE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <C2D7FE5348E1B147BCA15975FBA2307565BCA499@IN01WEMBXA.internal.synopsys.com>
Date:	Wed, 7 Jan 2015 06:36:14 +0000
From:	Vineet Gupta <Vineet.Gupta1@...opsys.com>
To:	Rob Herring <robh@...nel.org>
CC:	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	"linux-serial@...r.kernel.org" <linux-serial@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Peter Hurley <peter@...leysoftware.com>,
	Jiri Slaby <jslaby@...e.cz>, Arnd Bergmann <arnd@...db.de>
Subject: Re: [Patch v4] ARC: Dynamically determine BASE_BAUD from DeviceTree

On Tuesday 06 January 2015 08:08 PM, Rob Herring wrote:
> On Tue, Jan 6, 2015 at 7:59 AM, Vineet Gupta <Vineet.Gupta1@...opsys.com> wrote:
>> 8250 earlycon is broken on multi-platform ARC because the UART clk
>> value (BASE_BAUD) is fixed at build time.
> Note that it should only be broken if you rely on the kernel to init
> the uart. It should work if the boot loader configured the UART and
> you don't specify the baudrate.

But even if uboot set it up right - when the early 8250 is enabled in kernel, it
will try to apply BASE_BAUD to re init it again.
So if that doesn't match platform expectations, early prints will be garbled. Am I
missing something here?

>>  #include <asm/mach_desc.h>
>>
>> +#ifdef CONFIG_SERIAL_8250_CONSOLE
>> +
>> +static unsigned int arc_base_baud;
> This can be initdata.

OK !

>
>> +unsigned int __init arc_early_base_baud(void)
>> +{
>> +       return arc_base_baud/16;
>> +}
>> +
>> +static void __init arc_set_early_base_baud(unsigned long dt_root)
>> +{
>> +       unsigned int core_clk = arc_get_core_freq();
>> +
>> +       if (of_flat_dt_is_compatible(dt_root, "abilis,arc-tb10x"))
>> +               arc_base_baud = core_clk/3;
> How many platforms do you expect this to be? This scales to maybe 10,
> but not to 100 platforms. It certainly would not scale for ARM.

The need for this came from our internal development of 2 new platform based on
new ARCv2 ISA.
They are slated to hit mainline sometime this year. Hence this is in a sense prep
patch and converts the
only existing upstream user of BASE_BAUD for ARC (tb10x). For ARC atleast I don't
expect the scalability issue - yet :-)
ARM doesn't seem define asm/serial.h (BASE_BAUD) and it probably works OK with the
stub value in asm-generic ?


> If it
> is a lot, then we need to find a generic way to describe this in DT.
> For example, perhaps we require the uart node to have a
> clock-frequency property or add a chosen property. 

Yeah that would indeed be cleanest way. But I think we shd be ok for now.

> You could make this
> part of the machine descriptor instead, but that wouldn't be my first
> choice.

Me neither !

Thx,
-Vineet

>
> Rob
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/