[Fwd: [OLPC-devel] Flash driver status.]

[Ray.Tseng at quantatw.com]

-----Original Message-----
From: Ray Tseng (曾文瑞) 
Sent: Monday, May 15, 2006 10:20 AM
To: 'devel at laptop.org'
Subject: RE: [Fwd: [OLPC-devel] Flash driver status.]

David,
	We rely on the PIO setting in BIOS, and have a program to verify the NAND erase, program...
	The code in http://david.woodhou.se/olpc-enable-nand.c can set chip on MMIO, any sample program to verify the NAND erase, program ...?  

Ray Tseng 5/15/06

-----Original Message-----
From: David Woodhouse [mailto:dwmw2 at infradead.org]
Sent: Saturday, May 13, 2006 12:04 AM
To: Ray Tseng (曾文瑞)
Cc: Roger Huang (黃榮宗); nn at media.mit.edu; mark at laptop.org; walter at laptop.org; mlj at media.mit.edu; blizzard at redhat.com; jg at laptop.org
Subject: RE: [Fwd: [OLPC-devel] Flash driver status.]

On Fri, 2006-05-12 at 09:43 +0800, Ray.Tseng at quantatw.com wrote:
> 
> Enter CMOS setup: (Press F1 during POST) C. Motherboard Device 
> Configuration
> -> A. Drive Configuration
> -->Flash Configuration
> --->Flash Interface: Enabled
> --->Chip select 0 - size: 8K/16B
> --->Base: PCI default
> --->Type: NAND I/O
> 
> These procedure will allocate resources for NAND access.

OK, thanks. The 'Flash Interface:' option is well-hidden -- it doesn't actually get displayed until you cursor down to it, at which point it overwrites the 'Chip Select 0 - size:' option on the screen.

Whatever I set the base address to, it seems to be PIO instead of MMIO
-- even when I set it to 'D2000' which looks like it's intended to be a memory address.  We don't support PIO mode in the Linux driver; only MMIO -- is there any way to enable it in that mode?

The timing options (MSR_NANDF_DATA, MSR_NANDF_CTL) are set to their most conservative values too -- I get about 50% speed improvement if I set them to 0x01110111 and 0x00000111 respectively. 

I've updated http://david.woodhou.se/olpc-enable-nand.c accordingly.

The machine doesn't seem to like my Apple keyboard much. If I have it plugged in, it often doesn't see it -- and today it's also taken to getting hard drive errors. If I have only the (self-powered) USB hard drive and the (USB-powered) RTL8150 Ethernet connected, it's all fine.
Plug in just the keyboard and the hard drive though, and it won't boot
-- although it did yesterday.

I'd suspect power starvation but this lot works fine in Linux, and the power drain ought to be within spec. And it was working yesterday even when I had two mice plugged into the keyboard too.

Since most of my interaction with it (except when I'm poking at the
BIOS) is over the network, I'll just leave the keyboard unplugged for now. It may be more problematic for others though.

--
dwmw2