[PATCH RFC 0/3] arm64: disable/enable d-cache support for

[AKASHI Takahiro]

On 05/07/2015 01:54 PM, Pratyush Anand wrote:
>
>
> On Thursday 07 May 2015 09:55 AM, AKASHI Takahiro wrote:
>> Geoff, Pratyush
>>
>> On 05/07/2015 02:37 AM, Pratyush Anand wrote:
>>> Hi Geoff,
>>>
>>> On Wednesday 06 May 2015 10:35 PM, Geoff Levand wrote:
>>>> Your 2 minutes seems like a long time.  Is this seen on other platforms?
>>>
>>> At least it is seen on two different HW platforms I had tested with. I
>>> do not know about other.
>>>
>>> It would be nice if somebody else who is using kexec on HW platform
>>> can report their execution time.
>>
>> On my HW, it took about 42secs before showing extra boot messages after
>> "I'm in purgatory."
>>
>
> It would also depend on the size of different segments (mainly kernel and initrd) on which sha256 is calculated. What
> are the size of binaries in your case?

segment[0].memsz = 0x877000 // kernel
segment[1].memsz = 0x4000   // dtb
segment[2].memsz = 0x7000   // purgatory

no initrd under normal kexec.

> In my case:
>
> segment[0].memsz = 0x10e0000 // kernel
> segment[1].memsz = 0x1910000 // initrd
> segment[2].memsz = 0x10000   // dtb
> segment[3].memsz = 0x40000   // purgatory, not included in sha256
> segment[4].memsz = 0x10000   // elf core hdr

Your numbers have extra "0" :)

> @Geoff: Do you see any specific side effect for enabling D-cache in purgatory, other than the fact that it increases
> size of purgatory? May be we can keep an user option to select/not select d-cache enabling.

The discussions should be Cc'd to linux-arm-kernel (or at least, arm64 maintainers).

-Takahiro AKASHI