Re: [RFC PATCH kernel] powerpc: Fix early setup to make early_ioremap work

From: Christophe Leroy <christophe.leroy@csgroup.eu>
To: Alexey Kardashevskiy <aik@ozlabs.ru>, linuxppc-dev@lists.ozlabs.org
Subject: Re: [RFC PATCH kernel] powerpc: Fix early setup to make early_ioremap work
Date: Thu, 20 May 2021 07:38:59 +0200	[thread overview]
Message-ID: <f87ab395-e245-f09c-766e-87e2ccb2285c@csgroup.eu> (raw)
In-Reply-To: <74d06e55-62c3-6367-bb33-f6bbeffcac74@ozlabs.ru>

Le 20/05/2021 à 05:33, Alexey Kardashevskiy a écrit :
> Hm, my thunderbird says it is not cc:'ed but git sendmail says it did cc:
> 
> 
> Server: localhost
> MAIL FROM:<aik@ozlabs.ru>
> RCPT TO:<linuxppc-dev@lists.ozlabs.org>
> RCPT TO:<aik@ozlabs.ru>
> RCPT TO:<mpe@ellerman.id.au>
> RCPT TO:<christophe.leroy@csgroup.eu>
> From: Alexey Kardashevskiy <aik@ozlabs.ru>
> To: linuxppc-dev@lists.ozlabs.org
> Cc: Alexey Kardashevskiy <aik@ozlabs.ru>,
>          Michael Ellerman <mpe@ellerman.id.au>,
>          Christophe Leroy <christophe.leroy@csgroup.eu>
> Subject: [RFC PATCH kernel] powerpc: Fix early setup to make early_ioremap work
> 
> 
> Not sure what to believe.

I got the patch.

If you are looking at the mail you received from the ppc list, I think the list removes the members 
of the list from the Cc: in order to avoid the mail being received multiple times.

Christophe

> 
> 
> On 20/05/2021 13:29, Alexey Kardashevskiy wrote:
>> The immediate problem is that after
>> 0bd3f9e953bd ("powerpc/legacy_serial: Use early_ioremap()")
>> the kernel silently reboots. The reason is that early_ioremap() returns
>> broken addresses as it uses slot_virt[] array which initialized with
>> offsets from FIXADDR_TOP == IOREMAP_END+FIXADDR_SIZE ==
>> KERN_IO_END- FIXADDR_SIZ + FIXADDR_SIZE == __kernel_io_end which is 0
>> when early_ioremap_setup() is called. __kernel_io_end is initialized
>> little bit later in early_init_mmu().
>>
>> This fixes the initialization by swapping early_ioremap_setup and
>> early_init_mmu.
>>
>> This also fixes IOREMAP_END to use FIXADDR_SIZE defined just next to it,
>> seems to make sense, unless there is some weird logic with redefining
>> FIXADDR_SIZE as the compiling goes.
>>
>> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
>> ---
>>   arch/powerpc/include/asm/book3s/64/pgtable.h | 2 +-
>>   arch/powerpc/kernel/setup_64.c               | 3 ++-
>>   2 files changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h 
>> b/arch/powerpc/include/asm/book3s/64/pgtable.h
>> index a666d561b44d..54a06129794b 100644
>> --- a/arch/powerpc/include/asm/book3s/64/pgtable.h
>> +++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
>> @@ -325,8 +325,8 @@ extern unsigned long pci_io_base;
>>   #define  PHB_IO_END    (KERN_IO_START + FULL_IO_SIZE)
>>   #define IOREMAP_BASE    (PHB_IO_END)
>>   #define IOREMAP_START    (ioremap_bot)
>> -#define IOREMAP_END    (KERN_IO_END - FIXADDR_SIZE)
>>   #define FIXADDR_SIZE    SZ_32M
>> +#define IOREMAP_END    (KERN_IO_END - FIXADDR_SIZE)
>>   /* Advertise special mapping type for AGP */
>>   #define HAVE_PAGE_AGP
>> diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
>> index b779d25761cf..ce09fe5debf4 100644
>> --- a/arch/powerpc/kernel/setup_64.c
>> +++ b/arch/powerpc/kernel/setup_64.c
>> @@ -369,11 +369,12 @@ void __init early_setup(unsigned long dt_ptr)
>>       apply_feature_fixups();
>>       setup_feature_keys();
>> -    early_ioremap_setup();
>>       /* Initialize the hash table or TLB handling */
>>       early_init_mmu();
>> +    early_ioremap_setup();
>> +
>>       /*
>>        * After firmware and early platform setup code has set things up,
>>        * we note the SPR values for configurable control/performance
>>
> 