Re: [PATCH 0/3] ARM: Use generic interface to simplify crashkernel reservation

From: Jinjie Ruan <ruanjinjie@huawei.com>
To: Baoquan He <bhe@redhat.com>
Cc: <linux@armlinux.org.uk>, <vgoyal@redhat.com>, <dyoung@redhat.com>,
	<paul.walmsley@sifive.com>, <palmer@dabbelt.com>,
	<aou@eecs.berkeley.edu>, <arnd@arndb.de>, <afd@ti.com>,
	<akpm@linux-foundation.org>, <rmk+kernel@armlinux.org.uk>,
	<linus.walleij@linaro.org>, <eric.devolder@oracle.com>,
	<gregkh@linuxfoundation.org>, <deller@gmx.de>,
	<javierm@redhat.com>, <robh@kernel.org>,
	<thunder.leizhen@huawei.com>, <austindh.kim@gmail.com>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>, <kexec@lists.infradead.org>
Subject: Re: [PATCH 0/3] ARM: Use generic interface to simplify crashkernel reservation
Date: Wed, 10 Jul 2024 09:52:21 +0800	[thread overview]
Message-ID: <8aed5c59-090b-ea7e-020c-dd2785ee3d7e@huawei.com> (raw)
In-Reply-To: <Zo1D9sB0UTOIrApP@MiWiFi-R3L-srv>

On 2024/7/9 22:06, Baoquan He wrote:
> On 07/09/24 at 07:06pm, Jinjie Ruan wrote:
>>
>>
>> On 2024/7/9 18:39, Baoquan He wrote:
>>> On 07/09/24 at 05:50pm, Jinjie Ruan wrote:
>>>>
>>>>
>>>> On 2024/7/9 17:29, Baoquan He wrote:
>>>>> On 07/08/24 at 09:33pm, Jinjie Ruan wrote:
>>>>>> Currently, x86, arm64, riscv and loongarch has been switched to generic
>>>>>> crashkernel reservation. Also use generic interface to simplify crashkernel
>>>>>> reservation for arm32, and fix two bugs by the way.
>>>>>
>>>>> I am not sure if this is a good idea. I added the generic reservation
>>>>> itnerfaces for ARCH which support crashkernel=,high|low and normal
>>>>> crashkernel reservation, with this, the code can be simplified a lot.
>>>>> However, arm32 doesn't support crashkernel=,high, I am not sure if it's
>>>>> worth taking the change, most importantly, if it will cause
>>>>> misunderstanding or misoperation.
>>>>
>>>> Yes, arm32 doesn't support crashkernel=,high.
>>>>
>>>> However, a little enhancement to the generic code (please see the first
>>>> patch), the generic reservation interfaces can also be applicable to
>>>> architectures that do not support "high" such as arm32, and it can also
>>>> simplify the code (please see the third patch).
>>>
>>> Yeah, I can see the code is simplified. When you specified
>>> 'crashkernel=xM,high', do you think what should be warn out? Because
>>> it's an unsupported syntax on arm32, we should do something to print out
>>> appropriate message.
>>
>> Yes, you are right! In this patch it will print "crashkernel high memory
>> reservation failed." message and out for arm32 if you specify
> 
> That message may mislead people to believe crashkernel=,high is
> supported but reservation is failed, then a bug need be filed for this?
> We may expect a message telling this syntax is not supported on this
> ARCH.

"CRASH_ADDR_LOW_MAX >= CRASH_ADDR_HIGH_MAX" indicate that the arm32 does
not support "crashkernel=,high", I wonder if this is generic for similar
architecture. If so, the first patch can print such as
"crashkernel=,high is not supported on this ARCH" message.

> 
>> 'crashkernel=xM,high because "CRASH_ADDR_LOW_MAX" and
>> "CRASH_ADDR_HIGH_MAX" is identical for arm32. And it should also warn
>> out for other similar architecture.
>>
>>
>>>
>>>
>>
> 
> 

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec