Problem with MTD and multiple erase sizes

[Wolfgang Wegner]

Hi,

I am new to the list and also quite new to MTD, and - obviously - have
a problem.

I have MTD up and running with several partitions, however, I now have
a problem accessing the last partition which consists of a region with
a different erase size than the rest of the device.

Up to now, my algorithm is as follows:
- open device
- get info with MEMGETINFO
- loop over all erase blocks to do the unlocking/erase

This fails for the last partition because MEMGETINFO returns the erasesize
of 0x20000 for this partition, too.

I tried using MEMGETREGIONINFO, but this fails because MEMGETREGIONCOUNT
is 0 for this partition (most likely because there actually is only one
erasesize used within this partition).

Can anybody point me to what I am doing wrong here?

Thank you and best regards,
Wolfgang


Here is my kernel boot output:

ASTRO5373L flash device: 2000000 at 0
Number of erase regions: 2
Primary Vendor Command Set: 0001 (Intel/Sharp Extended)
Primary Algorithm Table at 010A
Alternative Vendor Command Set: 0000 (None)
No Alternate Algorithm Table
Vcc Minimum:  1.7 V
Vcc Maximum:  2.0 V
Vpp Minimum:  8.5 V
Vpp Maximum:  9.5 V
Typical byte/word write timeout: 256 µs
Maximum byte/word write timeout: 512 µs
Typical full buffer write timeout: 512 µs
Maximum full buffer write timeout: 1024 µs
Typical block erase timeout: 1024 ms
Maximum block erase timeout: 4096 ms
Chip erase not supported
Device size: 0x2000000 bytes (32 MiB)
Flash Device Interface description: 0x0001
  - x16-only asynchronous interface
Max. bytes in buffer write: 0x40
Number of Erase Block Regions: 2
  Erase Region #0: BlockSize 0x20000 bytes, 255 blocks
  Erase Region #1: BlockSize 0x8000 bytes, 4 blocks
ASTRO5373LFlash: Found 1 x16 devices at 0x0 in 16-bit bank
 Intel/Sharp Extended Query Table at 0x010A
 Intel/Sharp Extended Query Table at 0x010A
 Intel/Sharp Extended Query Table at 0x010A
 Intel/Sharp Extended Query Table at 0x010A
 Intel/Sharp Extended Query Table at 0x010A
Using buffer write method
cfi_cmdset_0001: Erase suspend on write enabled
Creating 9 MTD partitions on "ASTRO5373LFlash":
0x00000000-0x02000000 : "Complete Flash"
0x00000000-0x00080000 : "bootloader"
0x00080000-0x00280000 : "kernel"
0x00280000-0x00540000 : "root"
0x00540000-0x00800000 : "FPGA"
0x00800000-0x01f40000 : "data"
0x01f40000-0x01f80000 : "Parameters1"
0x01f80000-0x01fc0000 : "Parameters2"
0x01fc0000-0x02000000 : "Parameters_ext"



The partition in question is "Parameters_ext" which comprises 4 blocks
of 0x8000 bytes each.

For all my flash accesses I have some utility functions, here is the
unlock part (with some additional debug stuff already):

      if (ioctl(current_task.fd, MEMGETREGIONCOUNT, &region_count)) {
	fprintf(stderr,"flash_worker_threadfunc: MEMGETREGIONCOUNT(%d) failed. %d %s\n",
		current_task.fd, errno, strerror(errno)); 
	retval = -2;
	goto complete_fail;
      }
      region_info.regionindex = 0;
      if (ioctl(current_task.fd, MEMGETREGIONINFO, &region_info)) {
	fprintf(stderr,"flash_worker_threadfunc: MEMGETREGIONINFO(%d) failed. %d %s\n",
		current_task.fd, errno, strerror(errno)); 
	retval = -2;
	goto complete_fail;
      }
      fprintf(stderr, "flash_worker_threadfunc: regioninfo(%d): %d, 0x%x, 0x%x, 0x%x\n",
		current_task.fd,
		region_count,
                region_info.offset,
                region_info.erasesize,
                region_info.numblocks);
      if (ioctl(current_task.fd, MEMGETINFO, &meminfo)) {
	fprintf(stderr,"flash_worker_threadfunc: meminfo(%d) failed. %d %s\n",
		current_task.fd, errno, strerror(errno)); 
	retval = -2;
	goto complete_fail;
      }
      fprintf(stderr, "flash_worker_threadfunc: meminfo(%d): %d, 0x%x, 0x%x, 0x%x, 0x%x, 0x%x\n",
		current_task.fd,
		region_count,
                meminfo.type,
                meminfo.flags,
                meminfo.size,
                meminfo.erasesize,
                meminfo.writesize);

      if(BIT_IS_SET(current_task.command, FLASH_TASK_COMMAND_ERASE_BIT) ||
         BIT_IS_SET(current_task.command, FLASH_TASK_COMMAND_WRITE_BIT) ||
	 BIT_IS_SET(current_task.command, FLASH_TASK_COMMAND_UNLOCK_BIT)) {
	erase.start = current_task.base_address;
	erase.length = meminfo.erasesize;

	while(erase.start < address_end) {
	  fprintf(stderr,"flash_worker_thread: Performing Flash unlock on %d at offset 0x%x (size 0x%x)\n",
		  current_task.fd, erase.start, erase.length);
	  if (ioctl(current_task.fd, MEMUNLOCK, &erase)) {
	    fprintf(stderr,"flash_worker_thread: unlock failed. %d %s\n", errno, strerror(errno));
	    retval = -2;
	    goto complete_fail;
	  }
	  erase.start += meminfo.erasesize;
	}
      }

results in:
flash_worker_threadfunc: MEMGETREGIONINFO(4) failed. 22 Invalid argument

or

flash_worker_threadfunc: meminfo(4): 0, 0x3, 0xc00, 0x40000, 0x20000, 0x1
flash_worker_thread: Performing Flash unlock on 4 at offset 0x38000 (size 0x20000)
flash_worker_thread: unlock failed. 22 Invalid argument

in case I comment out the MEMGETREGIONINFO part.