OpenWrt Forum Archive

Topic: WRT54GL flash upgrade

The content of this topic has been archived on 23 Mar 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

Hi!

I upgraded my WRT54GL's flash with this chip: Spansion S29GL064A90TFIR40
http://www.newark.com/82K6613

I had to use a small wire to connect the additional address line (A21) from another footprint.

For details and pics, please look here:
http://www.dd-wrt.com/phpBB2/viewtopic. … 758#181758

This work also nicely with OpenWRT. (I have OpenVPN running on tis unit)

Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/root                 1024      1024         0 100% /rom
none                     31464        44     31420   0% /tmp
/dev/mtdblock/4           6336      1316      5020  21% /jffs
/jffs                     1024      1024         0 100% /

Note that I used tjtag-V2 to re-flash the CFE
http://www.dd-wrt.com/dd-wrtv2/down.php … do%2Fjtag/

elvicomte wrote:

And this one works too...

http://www.dd-wrt.com/phpBB2/viewtopic. … 163#182163

Cool smile
16MB flash is nice smile
I've been thinking about hooking up a NAND flash from and old 2Gb CF Card with 2x1Mx8bit NAND Flash Chips.
The IOs are surprisingly simple, I think adding the 8 data lines in parallel to the NOR Flash and using a bit of glue logic to generate new CS from
the NOR flash A2x and NOR flash CS for both NOR flash and NAND flash could work.
i.e. use part of the unused NOR flash address space as IO window for nand flash IO, CS and RE/WE and a few GPIOs for the remaining control signals
(ALE, CLE, RY/BY).  The driver part shouldn't be too difficult and (8 times wink) faster than the '1bit at a time' SD-Card mod.

I upgraded the flash ?n my WRT54GL 1.1,  with Spansion S29GL128P10TFI02 (16MB),using Kamikaze 8.09, 2.4 Kernel.
Now on every flash programming operation, I'm getting "Warning: DQ5 raised while program operation was in progress, however operation completed OK".
It is working, but some times, I'm getting random process kills, during package installation, and once the whole router was bricked, and I had to reflash it again.

I noticed similar messages with Whiterussian.
I wonder if this could be fixed by the way data is written to the chip???
Must be something in the CFE...

The discussion might have continued from here.