[PATCH/RFC 0/5] Another try for a new memtest implementation

[Sascha Hauer]

Hi Alex,

On Tue, Jun 04, 2013 at 01:13:47AM +0200, Alexander Aring wrote:
> Hi,
> 
> this is a new try to get a new memtest implementation into barebox.
> 
> I removed the ADDRESS_IN_REGIONS macro, which was a very ugly solution.

Indeed, much nicer now.

> The new memtest command used a list and save all unused sdram_regions
> of each bank. Then running the mem_test routine on these regions.
> After that the memtest command will release the requested sdram_regions.
> 
> I removed the start and stop command arguments to make the implementation
> easier/understandable and it's not necessary anyway.

Specifying a region would be nice, but not really necessary.

> 
> Maybe it's better to remove the silent ALIGN and ALIGN_DOWN from the start
> and end address in the common mem_test routine, because addresses are already
> align.
> It is difficult to explain because a end address has the form of 0xXXXXffff
> and a ALIGN_DOWN on it with a mask of 3 we have a calculated end address of
> 0xXXXXfffA which is wrong. (0xXXXXffff - 4) - 1...
> 
> Another question is:
> Can someone test it with more than one bank, please? :-)

I gave it a test on a karo tx25 board which has two memory banks. It
works just like expected :)

Overall this looks really good now. I just applied it to -next.

Thanks
 Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |