Usbat2 December 2004

usbat2@lists.lysator.liu.se

9 participants
32 discussions

USBAT 0.5.1 for Linux 2.6
by Daniel Drake 02 Jan '05

02 Jan '05

Hi, A new release - main changes being cleanups, annotations, and more sharing of code between usbat/usbat02. We also add support for the few USBAT02-based HP CD writers that exist in the wild. I fixed a few small issues with my own device too. ***WARNING*** You need to wipe out your kernel sources before applying this, upgrading from 0.5.0 will not work. Sorry, but this will probably be the last time that this needs to happen :) http://prdownloads.sourceforge.net/usbat2/usbat02-2.6-0.5.1.tar.gz?download All that is left to sort out is: 1. Rons slow writing 2. Rons problem with changing media (probably related) 3. Delays when changing media If you experience any of the above problems, please read on. I really need full debug logs to work on this. It would also be helpful if you could leave 60 seconds between each "operation", keeping an eye on the clock at the same time. This way you can say: At 12:01 I plugged in the card reader. At 12:02 I mounted it. At 12:03 I copied 100kb of files to the device At 12:04 I unmounted At 12:05 I unplugged This should make my reading through the logs be pretty clear. Ron, please take your problems one at a time! Ignoring the media changing for now, just do a process similar to the above, and send me the logs. I'd like to solve the writing problem before looking at the others. As for the delays with media changing - well a few of you said that you had to wait about 15 seconds for the "touch /dev/sda" operation to complete after changing media. This is not the case for me, it completes seemingly instantaneously. So I'd like to fix this up a bit more. Please perform a process similar to the above (at minute intervals: plug in card reader with media, change media, touch /dev/sda, unplug) and send the logs. Thanks for the testing of the last release, we're now pretty close to completion :) I'll be aiming to get us included for 2.6.11 Thanks! Daniel

4 4

USBAT02 0.5.0 for Linux 2.6
by Daniel Drake 27 Dec '04

27 Dec '04

Hi! Well, I recieved my 2nd device and have been having some fun with it. I've done quite a bit of work : previously, we were our own usb-storage "subdriver", but now, we are an addition to the already-existing shuttle_usbat driver. This makes a lot of sense, since unsuprisingly, the usbat02 does have some similarities to the original usbat :) Also, I have completed the driver feature-wise: you can now plug in the device with no media, then plug in the media, then mount the drive etc. Additionally, you can plug in some media, unplug the media, put in some other media, and then use the 2nd media with no problems at all. It's working great for me. (Microtech ZiO CF-II) There are now 3 items left on my todo list: - Wait for you to tell me it doesn't work at all - Fix it - Submit to mainline kernel There is one caveat to be aware of with respect to hotplugging/hotswapping media (this exists throughout usb-storage in general, not just our driver): When you change media, be sure to ask Linux to reread the partition table before trying to mount it. This can be done with a simple "touch /dev/sda". For example: 1. Plug in device with no media, sda appears 2. Plug in media, nothing happens 3. Run "touch /dev/sda", sda1 appears 4. mount sda1 Another example: 1. Plug in device with media, sda and sda1 appear 2. Unplug media, nothing happens optional --> 3. run "touch /dev/sda", sda1 disappears 4. Plug in other media, nothing happens 5. run "touch /dev/sda", sda1 appears 6. mount sda1 It has to be this way, because unfortunately, usb-storage devices don't send any signals at all to the host computer when media has changed! Windows gets around this by polling the device a few times each second, but Linux generally does not implement semi-redundant things like that. Another thing to be aware of - the usb-integration patch has changed a lot, and our code is now in shuttle_usbat.c (was previously shuttle_usbat02.c). To avoid any and all confusion, please start with a *clean* kernel. If you obtain the releases through CVS, please nuke your CVS checkout and do a fresh checkout. Then you can install the driver in the normal way. http://prdownloads.sourceforge.net/usbat2/usbat02-2.6-0.5.0.tar.gz?download I look forward to testing reports :) Thanks! Daniel

4 4

[Fwd: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to]
by Daniel Drake 24 Dec '04

24 Dec '04

-------- Original Message -------- Subject: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to Date: Fri, 24 Dec 2004 13:25:22 +0000 From: Daniel Drake <dsd(a)gentoo.org> To: Alan Stern <stern(a)rowland.harvard.edu> CC: autophile(a)gmail.com, linux-usb-devel(a)lists.sourceforge.net, mdharm-usb(a)one-eyed-alien.net References: <Pine.LNX.4.44L0.0412232216590.9167-100000(a)netrider.rowland.org> Alan Stern wrote: > Now you're confusing me. I thought these chips provided a USB-ATA > interface. Here you're talking about ISA registers. But ISA is a > motherboard architecture; it includes a lot more than just ATA. Does this > mean that the chips essentially provide an entire motherboard at the far > end of the USB link? Could you plug in an AT-style keyboard, for > instance? Yes, the USBAT chip seems quite a nifty one! You can connect ATA, EPP, and ISA devices to it. It's a pity that these things never really sold other than a few cd writers and CF readers - SCM have now dropped out of that market altogether... Here's a description of the chip: The shuttle USBAT-02 is a USB Function Interface ASIC providing easy integration of non-USB peripherals and peripheral controllers. The USBAT-02 is capable of attaching parallel port peripherals such as EPP scanners, IOMEGA ZIP drives and Shuttle parallel port peripherals. The USBAT-02 can also be configured to attach to direct ATA / ATAPI, ISA peripherals. Shuttle USBAT-02 is a single chip interface solution that requires no firmware or Micro Controller. > Don't ask me; I know zero about the HP8200. However the shuttle_usbat > driver does seem to use packet commands, according to the comments in the > source code. > > Doesn't the fact that the flash version returns an error indicate that > it really _isn't_ ATAPI? And doesn't your driver communicate with the > flash device using regular ATA commands, not ATAPI packet commands? Yes, these results certainly suggest this is the case. I had previously assumed that the flash devices were ATAPI since the compactflash spec says that CF devices are ATAPI-4 compliant. However now that I have a better idea about this, I guess its possible to be ATAPI-4 compliant and not implement packet commands. Either way, at least we have a possible distinction. I think I have enough to work on now. I'll try and find a HP8200 user to confirm this will work. Thanks a lot for all your help :) Daniel ------------------------------------------------------- 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/ _______________________________________________ linux-usb-devel(a)lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

1 0

[Fwd: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to]
by Daniel Drake 24 Dec '04

24 Dec '04

-------- Original Message -------- Subject: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to Date: Thu, 23 Dec 2004 22:30:31 -0500 (EST) From: Alan Stern <stern(a)rowland.harvard.edu> To: Daniel Drake <dsd(a)gentoo.org> CC: autophile(a)gmail.com, <linux-usb-devel(a)lists.sourceforge.net>, <mdharm-usb(a)one-eyed-alien.net> On Thu, 23 Dec 2004, Daniel Drake wrote: > The hp8200e init function does this: > > 730 // Write 0x80 to ISA port 0x3F > 731 > 732 if (usbat_write(us, USBAT_ISA, 0x3F, 0x80) != > 733 USB_STOR_XFER_GOOD) > 734 return USB_STOR_TRANSPORT_ERROR; > 735 > 736 US_DEBUGP("INIT 6\n"); > 737 > 738 // Read ISA port 0x27 > 739 > 740 if (usbat_read(us, USBAT_ISA, 0x27, status) != > 741 USB_STOR_XFER_GOOD) > 742 return USB_STOR_TRANSPORT_ERROR; > > The usbat02-flash init doesn't. I was hoping that reading/writing to those > registers on the flash device would fail (that would provide an easy > distinction) but unfortunately thats not the case. > > What do these reads/writes do? I did some searching but couldn't find any ISA > specs that didn't require me to pay for them. I have no idea. > On the flash-devices, ISA port 0x27 originally reads 0xEC. After performing > the same write as shown above, ISA 0x27 then reads 0x80. Now you're confusing me. I thought these chips provided a USB-ATA interface. Here you're talking about ISA registers. But ISA is a motherboard architecture; it includes a lot more than just ATA. Does this mean that the chips essentially provide an entire motherboard at the far end of the USB link? Could you plug in an AT-style keyboard, for instance? > Is there any sniffed data for the HP8200 available? Would anyone be available > to test a patch to spit out the values of these registers on the HP8200? I don't know. In fact, I don't know of anyone who has the HP8200. > Also, Alan, I did some reading up about ATA/ATAPI. If I understand it right, > ATA was only for hard disks, and ATAPI came later to support CD drives, etc. That's right. > This means that both the HP8200 and the flash devices are ATAPI. If you say so. But... > Interesting.. After executing IDENTIFY PACKET DEVICE (0xA1), the flash-usbat02 > returns an error (as indicated by bit 0 in status register) and does not > produce a high DRQ (bit 3 of status reg) as indicated in the command > description. So, the flash devices don't appear to support packet commands. Is > it safe to say that HP8200's *do* support these? Don't ask me; I know zero about the HP8200. However the shuttle_usbat driver does seem to use packet commands, according to the comments in the source code. Doesn't the fact that the flash version returns an error indicate that it really _isn't_ ATAPI? And doesn't your driver communicate with the flash device using regular ATA commands, not ATAPI packet commands? Alan Stern ------------------------------------------------------- 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/ _______________________________________________ linux-usb-devel(a)lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

1 0

GNOME experience
by Daniel Drake 24 Dec '04

24 Dec '04

Hi, I got gnome/dbus/hal setup. I was unable to get any image import wizards popping up when i plugged the device in, but I had a play around (it was automounting the usbat and putting an icon on my desktop on plugin). It uses the options in the fstab, and you do need to unmount it manually (right click on icon --> unmount volume). So it looks like gnome isn't quite ready for auntie tillie yet :) Lindsay, perhaps you could test 0.5.0 and let me know if it still gives you your dbus/hal hanging issue. Thanks, Daniel

2 3

[Fwd: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to]
by Daniel Drake 24 Dec '04

24 Dec '04

-------- Original Message -------- Subject: Re: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to Date: Thu, 23 Dec 2004 23:39:40 +0000 From: Daniel Drake <dsd(a)gentoo.org> To: Daniel Drake <dsd(a)gentoo.org> CC: Alan Stern <stern(a)rowland.harvard.edu>, autophile(a)gmail.com, linux-usb-devel(a)lists.sourceforge.net, mdharm-usb(a)one-eyed-alien.net References: <Pine.LNX.4.44L0.0412222252490.28573-100000(a)netrider.rowland.org> <41CAB653.6040207(a)gentoo.org> Something I wrote in my last mail gave me an idea.. Daniel Drake wrote: > In this case, both the hp8200 and the flash devices use the ATA access > almost entirely (the hp8200 writes to some ISA registers during the init > function, not really sure why..). The hp8200e init function does this: 730 // Write 0x80 to ISA port 0x3F 731 732 if (usbat_write(us, USBAT_ISA, 0x3F, 0x80) != 733 USB_STOR_XFER_GOOD) 734 return USB_STOR_TRANSPORT_ERROR; 735 736 US_DEBUGP("INIT 6\n"); 737 738 // Read ISA port 0x27 739 740 if (usbat_read(us, USBAT_ISA, 0x27, status) != 741 USB_STOR_XFER_GOOD) 742 return USB_STOR_TRANSPORT_ERROR; The usbat02-flash init doesn't. I was hoping that reading/writing to those registers on the flash device would fail (that would provide an easy distinction) but unfortunately thats not the case. What do these reads/writes do? I did some searching but couldn't find any ISA specs that didn't require me to pay for them. On the flash-devices, ISA port 0x27 originally reads 0xEC. After performing the same write as shown above, ISA 0x27 then reads 0x80. Is there any sniffed data for the HP8200 available? Would anyone be available to test a patch to spit out the values of these registers on the HP8200? Also, Alan, I did some reading up about ATA/ATAPI. If I understand it right, ATA was only for hard disks, and ATAPI came later to support CD drives, etc. This means that both the HP8200 and the flash devices are ATAPI. Thanks, Daniel ------------------------------------------------------- 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/ _______________________________________________ linux-usb-devel(a)lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

1 0

[Fwd: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to]
by Daniel Drake 24 Dec '04

24 Dec '04

-------- Original Message -------- Subject: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to Date: Fri, 24 Dec 2004 00:27:37 +0000 From: Daniel Drake <dsd(a)gentoo.org> To: Alan Stern <stern(a)rowland.harvard.edu> CC: autophile(a)gmail.com, linux-usb-devel(a)lists.sourceforge.net, mdharm-usb(a)one-eyed-alien.net References: <Pine.LNX.4.44L0.0412231730160.19055-100000(a)netrider.rowland.org> Alan Stern wrote: > Well, the IDENTIFY PACKET DEVICE command looks relevant. Beyond that, you > need to rely on appropriate packet commands. INQUIRY is implemented as a > packet command; it might be the one you need. Interesting.. After executing IDENTIFY PACKET DEVICE (0xA1), the flash-usbat02 returns an error (as indicated by bit 0 in status register) and does not produce a high DRQ (bit 3 of status reg) as indicated in the command description. So, the flash devices don't appear to support packet commands. Is it safe to say that HP8200's *do* support these? Thanks, Daniel ------------------------------------------------------- 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/ _______________________________________________ linux-usb-devel(a)lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

1 0

[Fwd: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to]
by Daniel Drake 24 Dec '04

24 Dec '04

-------- Original Message -------- Subject: [linux-usb-devel] Re: Detecting which kind of device the USBAT is connected to Date: Thu, 23 Dec 2004 17:37:08 -0500 (EST) From: Alan Stern <stern(a)rowland.harvard.edu> To: Daniel Drake <dsd(a)gentoo.org> CC: autophile(a)gmail.com, <linux-usb-devel(a)lists.sourceforge.net>, <mdharm-usb(a)one-eyed-alien.net> On Thu, 23 Dec 2004, Daniel Drake wrote: > I'm fairly new to this and am not too sure on the distinction between ATA and > ATAPI. I'll have to do some reading up, but thanks for the suggestion :) ATA was the original protocol used for hard disk controllers. ATAPI was an extension that added support for "Packet" commands. Hence the name: ATA plus "Packet Interface". These used the ATA hardware but were a completely different command set, basically duplicating the SCSI commands sets. The intent was to allow cdrom drives (which were originally designed as SCSI devices) to work over the IDE (= ATA) interface. > > Again, I'm stymied without the specs. Is there some ATA-level command you > > can use to find out whether a device supports ATAPI? > > I'm not sure, but I don't think this would help much: both the hp8200 and the > flash readers use the ATA/ATAPI transport. Or are you suggesting that the > hp8200 implements ATA but not ATAPI? Well, the IDENTIFY PACKET DEVICE command looks relevant. Beyond that, you need to rely on appropriate packet commands. INQUIRY is implemented as a packet command; it might be the one you need. Alan Stern ------------------------------------------------------- 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/ _______________________________________________ linux-usb-devel(a)lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

1 0

Re: USBAT02 0.5.0 for Linux 2.6
by Daniel Drake 23 Dec '04

23 Dec '04

Hi Ron, Ron wrote: > Tried changing the media card. umount, then touch /dev/sda worked as you > said it should. Plugged in a 256 Mb card instead of the 4 Mb card and > did touch /dev/sda again. It hung for a very long time, (had a shower, > got dressed) then failed with unable to read partition table. I think I > tried this after I had already copied messages, so probably not in this > copy. I booted back into 2.4 kernel and the card reads fine, no data lost. > It also crashes Konqueror, but this may be a Mandrake bug. There is an > upgrade for a problem where plugging in a USB memory stick crashes > Konqueror. I'm on dial up so can't really try the update to see if it > helps. Unplug the drive, rmmod usb-storage, init 3, kill -9 something > called "kdeinit:kded" then init 5 and Konqueror comes back to life. Actually, this is actually crashing your kernel. This may be related to the problems you are having - the kernel is half-crashed, it might not be working as expected. Either way, I really need to see debug logs to investigate this. Could you please compile your kernel with USB debugging _and_ USB storage debugging enabled and send me complete logs? If you are able to apply the crashing fix without too much trouble, that would also be appreciated. Which USBAT device is it that you are running? By the way, I noticed that your hdb disk drive is reporting some errors. If you aren't aware of this, might be worth using a SMART tool to see whats up. Thanks, Daniel

1 0

0.5.0 Closer Still
by Kurt Wall 23 Dec '04

23 Dec '04

Hi, list, With a properly configured kernel that actually has the 0.5.0 driver, detection and device creation works: --- /var/log/messages snippet --- Dec 23 08:53:40 luther kernel: usb 1-1: new full speed USB device using address 2 Dec 23 08:53:41 luther kernel: usb-storage: USBAT-02 Version 0.5.0, init_usbat02 called. Dec 23 08:53:44 luther kernel: scsi0 : SCSI emulation for USB Mass Storage devices Dec 23 08:53:44 luther kernel: Vendor: SCM Model: SCM USBAT-02 Rev: 0005 Dec 23 08:53:44 luther kernel: Type: Direct-Access ANSI SCSI revision: 02 Dec 23 08:53:44 luther scsi.agent[2303]: how to add device type= at /devices/pci0000:00/0000:00:07.2/usb1/1-1/1-1:1.0/host0/0:0:0:0 ?? Dec 23 08:53:44 luther kernel: SCSI device sda: 32128 512-byte hdwr sectors (16 MB) Dec 23 08:53:44 luther kernel: sda: assuming Write Enabled Dec 23 08:53:44 luther kernel: sda: sda1 Dec 23 08:53:44 luther kernel: Attached scsi removable disk sda at scsi0, channel 0, id 0, lun 0 Dec 23 08:53:44 luther kernel: Attached scsi generic sg0 at scsi0, channel 0, id 0, lun 0, type 0 Dec 23 08:53:45 luther udev[2323]: creating device node '/dev/sda' Dec 23 08:53:45 luther udev[2331]: creating device node '/dev/sda1' Dec 23 08:53:45 luther udev[2347]: creating device node '/dev/sg0' --- /var/log/messages snippet --- Mounting complains about a bad superblock: root@luther:~# mount -t vfat /dev/sda1 /mnt/cf mount: wrong fs type, bad option, bad superblock on /dev/sda1, or too many mounted file systems I assumed that the filesystem on the card was vfat. However, fdisk revealed otherwise: root@luther:~# fdisk -l /dev/sda Disk /dev/sda: 16 MB, 16449536 bytes 4 heads, 32 sectors/track, 251 cylinders Units = cylinders of 128 * 512 = 65536 bytes Device Boot Start End Blocks Id System /dev/sda1 * 1 250 15984 1 FAT12 FAT12 is MS-DOS, so I retried with -t msdos, and it worked: root@luther:~# ls -l /mnt/cf/dcim/100hp210 -rwxr--r-- 1 root root 292857 2004-11-21 13:51 im000811.jpg* -rwxr--r-- 1 root root 322199 2004-11-21 16:45 im000812.jpg* -rwxr--r-- 1 root root 339972 2004-11-21 16:45 im000813.jpg* -rwxr--r-- 1 root root 356969 2004-11-21 17:05 im000814.jpg* -rwxr--r-- 1 root root 380617 2004-11-21 17:06 im000815.jpg* So, to make a long story short: Thanks, Daniel, for all of your work! It will be great to have this working for the upcoming Christmas festivities I was surpised that the camera put a FAT12 filesystem on the card, but at least I can work with that. Thanks again. Kurt -- Ducharme's Precept: Opportunity always knocks at the least opportune moment.

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

Usbat2 December 2004