Hi. I am trying to get the ohci-isp1362 driver working, and I feel like I am almost there. Here's my situation:
The only thing I'm using it for right now is to transfer files to and from a flash drive. I can enumerate fine, and it seems I can list files and directories just fine too. I can write files, if I write them 4KiB at a time, and 'sync' between writes. Reading sometimes fails, and it seems if I read straight to SDRAM it fails more often than reading to onboard NAND flash (JFFS2). I'm assuming this has to do with various timing issues. I'm trying to analyze my debug messages, and understand what's happening. Could you give some guidance? Here's the pattern that I'm seeing in my debug messages: (note: My debug messages likely look a bit different) ohci_write_control: HCCONTROL 00000083->000000a3 In 'td_submit_urb'. ohci_write_cmdstatus: HCCMDSTAT 00000000->00000004 find_td: Found ED c10f0080(c10f0080)... td_flags:f0100000 Processing TD c1080040 ED c10f0080 PTD[0] 0610 #1010 enqueue_ptd: Queueing PTD 0610 len 512 TD c1080040 #1010 __isp1362_write_buffer: Writing 8 byte to buffer[00] @ 0610 __isp1362_read_buffer: Reading 72 byte from buffer[00] @ 0610 flush_ptd: Flushing PTD[0] 0610, TD c1080040 #102d __isp1362_read_buffer: Reading 8 byte from buffer[00] @ 0610 __isp1362_read_buffer: Reading 512 byte from buffer[00] @ 0618 Retiring TD c1080040 ED c10f0080 info 00401102 with CC=0 #102d finish_xfer: PTD 0610 of TD c1080040 finished #102d ohci_write_control: HCCONTROL 000000a3->00000083 Now, 0x0610 is the area in the buffer for the first ATL packet. It can handle 512 bytes, and it is transferred on the bus in 64 byte chunks. Now, the irregularity (and trouble) starts with these debug messages: ohci_write_control: HCCONTROL 00000083->000000a3 In 'td_submit_urb'. ohci_write_cmdstatus: HCCMDSTAT 00000000->00000004 find_td: Found ED c10f0080(c10f0080)... td_flags:f0d00000 Processing TD c1080040 ED c10f0080 PTD[0] 0610 #10eb enqueue_ptd: Queueing PTD 0610 len 512 TD c1080040 #10eb __isp1362_write_buffer: Writing 8 byte to buffer[00] @ 0610 __isp1362_read_buffer: Reading 72 byte from buffer[00] @ 0610 In 'td_submit_urb'. ohci_write_cmdstatus: HCCMDSTAT 00000000->00000004 flush_ptd: Flushing PTD[0] 0610, TD c1080040 #1108 __isp1362_read_buffer: Reading 8 byte from buffer[00] @ 0610 __isp1362_read_buffer: Reading 512 byte from buffer[00] @ 0618 - In 'requeue_ptd' enqueue_ptd: Queueing PTD 0610 len 512 TD c1080040 #1108 __isp1362_write_buffer: Writing 8 byte to buffer[00] @ 0610 __isp1362_read_buffer: Reading 72 byte from buffer[00] @ 0610 find_td: Found ED c10f0080(c10f0080)... td_flags:a3d00000 _!_ td = td->hwNextTD (Break here? -emu.c:~904) td_flags:f0d00000 Processing TD c1080000 ED c10f0080 PTD[1] 0818 #1134 enqueue_ptd: Queueing PTD 0818 len 512 TD c1080000 #1134 __isp1362_write_buffer: Writing 8 byte to buffer[00] @ 0818 __isp1362_read_buffer: Reading 72 byte from buffer[00] @ 0818 Now, I'm in the middle of processing a TD, and I get these lines that don't normally happen here: In 'td_submit_urb'. ohci_write_cmdstatus: HCCMDSTAT 00000000->00000004 Is this because it's processing the next TD while waiting for the first one? I also see a new 'find_td' message, but I never see the 'Retiring TD' message for the first one. This then causes it to go on to the next PTD buffer area (@ 0x0818). It then keeps on doing the same thing, and interrupts with a new 'find_td' with the next PTD buffer area after that. It finally halts with the "update_td_status: TD c1080040 ERROR 3 TD_CC_DATATOGGLEM" error.. I'm assuming since it missed something, it didn't correctly set the toggle bit. Now, is something happening concurrently that's adding new TDs before the last one has finished if it's not fast enough? Could I add some sort of gate to prevent processing of a TD before the old one has finished? (I also noticed the '0xa' at the start of td_flags, which is the TD_BUSY condition code added by the isp's ohci emulation, and it doesn't seem to ever get cleared) I really don't understand how this all works, or what's going on, so any pointers / insight are more than welcome. Thanks, Mike ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://productguide.itmanagersjournal.com/ _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
