[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87jzhdkask.fsf@BLaptop.bootlin.com>
Date: Mon, 22 Jul 2024 11:58:51 +0200
From: Gregory CLEMENT <gregory.clement@...tlin.com>
To: Jiaxun Yang <jiaxun.yang@...goat.com>, Thomas Bogendoerfer
<tsbogend@...ha.franken.de>, "paulburton@...nel.org"
<paulburton@...nel.org>
Cc: "linux-mips@...r.kernel.org" <linux-mips@...r.kernel.org>,
linux-kernel@...r.kernel.org, Théo Lebrun
<theo.lebrun@...tlin.com>,
Vladimir Kondratiev <vladimir.kondratiev@...ileye.com>, Tawfik Bayouk
<tawfik.bayouk@...ileye.com>, Thomas Petazzoni
<thomas.petazzoni@...tlin.com>
Subject: Re: [PATCH] MIPS: SMP-CPS: Fix address for GCR_ACCESS register for
I6500
Hello Jiaxun,
> 在2024年7月20日七月 上午11:13,Jiaxun Yang写道:
>> 在2024年7月19日七月 下午10:14,Gregory CLEMENT写道:
>>> Unlike most other MIPS CPUs, the I6500 CPUs have different address
>>> offsets for the Global CSR Access Privilege register. In the "MIPS64
>>> I6500 Multiprocessing System Programmer's Guide," it is stated that
>>> "the Global CSR Access Privilege register is located at offset 0x0120"
>>> in section 5.4.
>>>
>>> However, this is not the case for other MIPS64 CPUs such as the
>>> P6600. In the "MIPS64® P6600 Multiprocessing System Software User's
>>> Guide," section 6.4.2.6 states that the GCR_ACCESS register has an
>>> offset of 0x0020.
>>
>> Hi Gregory,
>>
>> I confirmed this is a CM3 feature rather than CPU core (Samruai) feature.
>
> Oh I’m not really sure if it’s CM 3.5 only.
>
> Let me check this Monday once I can checkout old design database for
> I6400.
Ok, so I am waiting for your feedback :)
And I am also trying to see if I can find the datasheet for I6400.
>
> Hardware resets GCR_ACCESS to the most permissive value so I assume
> it’s your bootloader doing wired hacks.
Indeed, other bootloaders seem to not modify it, so that's why I think
the issue was never detected until now. However, we want to be as
independent as possible from the bootloader, so we really need to fix
it.
Thanks for your review!
Gregory
>
> Thanks
>
>>
>> Please use CM version to select register region.
>> (And perhaps Cc stable for this patch?)
>>
>> Thanks
>> - Jiaxun
>>
>>>
>>> This fix allows to use the VP cores in SMP mode.
>>>
>>> Based on the work of Vladimir Kondratiev <vladimir.kondratiev@...ileye.com>
>>>
>>> Fixes: 859aeb1b0dd1 ("MIPS: Probe the I6500 CPU")
>>> Signed-off-by: Gregory CLEMENT <gregory.clement@...tlin.com>
>>> ---
>>> arch/mips/include/asm/mips-cm.h | 4 ++++
>>> arch/mips/kernel/smp-cps.c | 5 ++++-
>>> 2 files changed, 8 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/mips/include/asm/mips-cm.h b/arch/mips/include/asm/mips-cm.h
>>> index 3d9efc802e36..41bf9b3a98fb 100644
>>> --- a/arch/mips/include/asm/mips-cm.h
>>> +++ b/arch/mips/include/asm/mips-cm.h
>>> @@ -240,6 +240,10 @@ GCR_ACCESSOR_RO(32, 0x0d0, gic_status)
>>> GCR_ACCESSOR_RO(32, 0x0f0, cpc_status)
>>> #define CM_GCR_CPC_STATUS_EX BIT(0)
>>>
>>> +/* GCR_ACCESS - Controls core/IOCU access to GCRs */
>>> +GCR_ACCESSOR_RW(32, 0x120, access_i6500)
>>> +#define CM_GCR_ACCESS_ACCESSEN GENMASK(7, 0)
>>> +
>>> /* GCR_L2_CONFIG - Indicates L2 cache configuration when Config5.L2C=1
>>> */
>>> GCR_ACCESSOR_RW(32, 0x130, l2_config)
>>> #define CM_GCR_L2_CONFIG_BYPASS BIT(20)
>>> diff --git a/arch/mips/kernel/smp-cps.c b/arch/mips/kernel/smp-cps.c
>>> index e074138ffd7f..60590890b6da 100644
>>> --- a/arch/mips/kernel/smp-cps.c
>>> +++ b/arch/mips/kernel/smp-cps.c
>>> @@ -325,7 +325,10 @@ static void boot_core(unsigned int core, unsigned
>>> int vpe_id)
>>> write_gcr_co_reset_ext_base(CM_GCR_Cx_RESET_EXT_BASE_UEB);
>>>
>>> /* Ensure the core can access the GCRs */
>>> - set_gcr_access(1 << core);
>>> + if (current_cpu_type() != CPU_I6500)
>>> + set_gcr_access(1 << core);
>>> + else
>>> + set_gcr_access_i6500(1 << core);
>>>
>>> if (mips_cpc_present()) {
>>> /* Reset the core */
>>>
>>> ---
>>> base-commit: 9298d51eb3af24f88b211087eb698399f9efa439
>>> change-id: 20240719-smp_i6500-8cb233878c41
>>>
>>> Best regards,
>>> --
>>> Gregory CLEMENT <gregory.clement@...tlin.com>
>>
>> --
>> - Jiaxun
>
> --
> - Jiaxun
Powered by blists - more mailing lists