Nope. News to me of its existence. Have any idea of the ub driver's size?
The trick is to get at least restore from USB storage able to be in the boot flash; I don't mind so much if we have to put TCP/IP and wireless modules into a partition on the regular flash. And if we can get it all to fit into less than 512K bytes (BIOS + Linux as bootloader), then we get to save $.3 to $.5, which over the number of units we're building, is more than small change. Thanks, - Jim
On Wed, 2006-03-15 at 08:45 -0500, Alan Cox wrote:
On Wed, Mar 15, 2006 at 07:22:57AM -0500, Jim Gettys wrote:
The biggest headache we have here is the size of the USB modules for usb storage and wireless,
and the drivers themselves. This is why we're
Have you looked at Pete Zaitcev's 'ub' driver instead of the full scsi stack ?
-- olpc-software mailing list olpc-software@redhat.com https://www.redhat.com/mailman/listinfo/olpc-software
On Wed, Mar 15, 2006 at 08:55:36AM -0500, Jim Gettys wrote:
Nope. News to me of its existence. Have any idea of the ub driver's size?
On x86-64 (no legacy 32bit box to hand)
15032 1400 28 16460 404c /lib/modules/2.6.9-ac3/kernel/drivers/block/ub.ko
If you are usign CF it will also be possible to do a dedicated PIO CF driver a similar way to get rid of the IDE layer or libata. Wouldn't let you do CD tho.
Symbol set: usb_deregister unregister_blkdev usb_register register_blkdev blk_cleanup_queue del_gendisk _spin_unlock_irq _spin_lock_irq usb_put_dev device_remove_file put_disk add_disk blk_max_low_pfn blk_queue_max_sectors blk_queue_max_phys_segments blk_queue_max_hw_segments blk_init_queue alloc_disk device_create_file usb_get_dev snprintf tasklet_init usb_init_urb del_timer_sync wait_for_completion kmem_cache_alloc malloc_sizes complete check_disk_change usb_unlink_urb del_timer usb_submit_urb __mod_timer jiffies memcpy __tasklet_schedule end_that_request_last end_that_request_first blk_start_queue blk_stop_queue elv_remove_request elv_next_request kfree printk sprintf _spin_unlock_irqrestore _spin_lock_irqsave