linux-kernel - Re: [PATCH 5/6] ARM: convert PCI defines to variables

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]

Date:	Sat, 02 Jul 2011 07:40:19 -0500
From:	Rob Herring <robherring2@...il.com>
To:	Arnd Bergmann <arnd@...db.de>
CC:	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	Michal Simek <monstr@...str.eu>, nico@...xnic.net,
	Rob Herring <rob.herring@...xeda.com>
Subject: Re: [PATCH 5/6] ARM: convert PCI defines to variables

Arnd,

On 07/02/2011 04:21 AM, Arnd Bergmann wrote:
> Hi Rob,
> 
> On Wednesday 29 June 2011 18:46:58 Rob Herring wrote:
>> From: Rob Herring <rob.herring@...xeda.com>
>>
>> Convert PCIBIOS_MIN_IO and PCIBIOS_MIN_MEM to variables to allow
>> multi-platform builds. This also removes the requirement for a platform to
>> have a mach/hardware.h.
> 
> Good idea.
> 
>> @@ -24,6 +23,12 @@ EXPORT_SYMBOL(ioport_unmap);
>>  #endif
>>  
>>  #ifdef CONFIG_PCI
>> +unsigned long pcibios_min_io;
>> +EXPORT_SYMBOL(pcibios_min_io);
>> +
>> +unsigned long pcibios_min_mem;
>> +EXPORT_SYMBOL(pcibios_min_mem);
>> +
> 
> How about setting these to sensible defaults that will work on
> most platforms, so that the majority won't have to set them at
> all?
> 
> Basically, the defaults should clearly be 
> 
> unsigned long pcibios_min_io  = 0x1000;
> unsigned long pcibios_min_mem = 0x01000000;
> 
> This just gets us out of the ISA bus range, so an ISA card behind
> a bridge can use all cards correctly. Most of the ones that
> currently set both to zero can probably just use those defaults
> as well, but some architectures have multiple buses or don't start
> the memory range at zero, so they might need higher values.

For many platforms, the defaults are 0 for both of these, so I went with
that. Do you think those are just wrong?

Rob

> 
>> --- a/arch/arm/mach-versatile/include/mach/hardware.h
>> +++ b/arch/arm/mach-versatile/include/mach/hardware.h
>> @@ -30,10 +30,6 @@
>>  #define VERSATILE_PCI_VIRT_BASE		(void __iomem *)0xe8000000ul
>>  #define VERSATILE_PCI_CFG_VIRT_BASE	(void __iomem *)0xe9000000ul
>>  
>> -/* CIK guesswork */
>> -#define PCIBIOS_MIN_IO			0x44000000
>> -#define PCIBIOS_MIN_MEM			0x50000000
>> -
>>  /* macro to get at IO space when running virtually */
>>  #define IO_ADDRESS(x)		(((x) & 0x0fffffff) + (((x) >> 4) & 0x0f000000) + 0xf0000000)
>>  
> 
> This PCIBIOS_MIN_IO setting is wrong, and PIO doesn't work on versatile
> because of this. I have an older patch series that I should dig out again
> to fix them and make versatile use the defaults. Don't worry about this
> one.
> 
> 	Arnd

--
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/