Re: Problem accessing all available NOR flash memory running uCLinux

Altera_Forum · ‎10-23-2014

Hi,

I kindly request for your expertise to help me with this issue. I have been stuck with this problem about couple weeks already and have search and tried multiple suggestions but still no luck. This includes tried with the latest Linux driver. I have a custom board with a 128MByte NOR flash memory 16 bit databus Part:PC28F00AP33TFA, and am having trouble to access the entire memory. The uClinux was able to detect the NOR flash memory by printing this message: "20000000.flash: Found 1 x32 devices at 0x0 in 16-bit bank. Manufacturer ID 0x000089 Chip ID 0x008966"

Intel/Sharp Extended Query Table at 0x010A

Using buffer write method

Using auto-unlock on power-up/resume

cfi_cmdset_001: Erase suspend on write enabled

erase region 0: offset=0x0,size=0x20000,blocks=1023

erase region 1: offset=0x7fe0000,size=0x8000,blocks=4

I know it know it understand there is 128MB because when I "cat /proc/mtd", the mtd0 size reported back is 08000000, which about 134MB, and detected erasesize is 00020000. However, when everything seems to be okay, I tried to "flash_eraseall -j /dev/mtd0", I get error(s):

Erasing 128Kibyte@ 0 -0% complete.20000000.flash: buffer write error (status 0xb0)

flash_eraseall: /dev/mtd0: short write

Not to mention, I can't mount my chip to mtdblock0:( I am using device tree (dts) configuration and am not sure if I missed anything?

The following is my setup:

In Qsys "Generic Tri-State Controller" setting: address width:27, data width:16, Byteenable width:2, Byte per word:2

Top level HDL:

output wire [25:0] flash_address,

inout reg [15:0] flash_data,

.

wire [26:0] addr_internal;

assign flash_address = addr_internal[26:1];

In uClinux menuconfig:

Device driver--> <*>Memory Technology Device (MTD) support-->

<*> Command line partition table parsing

<*>Openfirmware partitioning information support

<*>Direct char device access to MTD devices

-*- Common interface to block layer for MTD 'translation layers'

<*> Caching block device access to MTD devices

RAM/ROM/Flash chip drivers-->

<*> Detect flash chips by Common Flsah Interface (CFI) probe

[*] Flash chip driver advanced configuration options (Flash cmd/query data swapping (NO)

[*] Specific CFI Flash gemometry selection

[*] Support 8-bit buswidth

[*] Support 16-bit buswidth

[*] Support 1-chip flash interleave

[*] Support 2-chip flash interleave

<*> Support for Intel/Sharp flash chips

Mapping drivers for chip access -->

<*> Flash device in physical memory map based on OF description //I disabled the Flash device in physical memory map

In device tree:

flash: flash@0x0 {

compatible ="ALTR,cfi_flash-13.1", "cfi-flash";

reg =< 0x00000000 0x08000000 >;

bank-width = < 2 >;

device-width = < 1 >;

};

The NOR memory works only if I access half of the memory (67MByte). No erase error or read/write problem, and mounted successful. The only different is configuration of the HDL address bus is 26 bits instead of 27 as following:

In Qsys "Generic Tri-State Controller" setting: address width:26, data width:16, Byteenable width:2, Byte per word:2

Top level HDL:

output wire [25:0] flash_address,

inout reg [15:0] flash_data,

.

wire [25:0] addr_internal;

assign flash_address = {1'b0, addr_internal[25:1]};

Any help is greatly appreciated:)

Altera_Forum · ‎10-27-2014

Hi,

Days have gone by and I had tried several things but still no luck of resolving this issue. With this part I'm seeing MTD errors that I think I've traced to cfi_cmdset_0001.c that I'd like to ask about. I removed the commented CMDSET0001_DISABLE_ERASE_SUSPEND_ON_WRITE and CMDSET0001_DISABLE_WRITE_SUSPEND in cfi_cmdset_0001.c, but the result is the same "buffer write error (status 0xb0)". According to the datasheet, 0xb0 means program or erase suspend. I am able to flash_eraseall /dev/mtd0 without the "-j" clean marker, but when I tried to flash_eraseall -j /dev/mtd0, it gives me the same error. why?