From: Douglas G. <do...@to...> - 2008-06-04 02:09:34
|
Bruce Allen wrote: > Could someone from the developers list respond to this? > > ---------- Forwarded message ---------- > Date: Sun, 25 May 2008 12:47:20 -0700 > From: Jim Kleckner <jek...@kl...> > To: sma...@li... > Subject: [smartmontools-support] smartctl USB support (atacb) > > The recent patches for atacb support for USB drives > with Cypress chipsets are exciting. > > Very often you are trying to deal with drives using a > USB2.0 bridge like the one from Vantec. > > Can you describe how we can tell whether smartctl > will work with some form of pass-through for a given > hardware configuration? > > How can we tell if there is a Cypress chipset? > > A large number of motherboards use the Intel chipsets. > Will there ever be a way to use those USB systems > with smartctl? I went looking in Linux for any indication what the USB storage class bridge device is and found nothing. I doubt that the usb-storage driver in the Linux kernel knows, but if it does, I can find no report. Perhaps the driver maintainer may correct me. Doug Gilbert [cc-ed linux usb-storage driver maintainer] |
From: Bruce A. <ba...@gr...> - 2008-06-09 14:45:24
|
> But you may check the usb.ids table. > > This one worked with the latest smartmontools: > ID 04b4:6830 Cypress Semiconductor Corp. CY7C68300A EZ-USB AT2 USB 2.0 to ATA/ATAPI Where does one get this data? What command(s) are used to get it? Cheers, Bruce |
From: Stanislav B. <sb...@su...> - 2008-06-09 15:27:11
|
Bruce Allen wrote: > > But you may check the usb.ids table. > > > > This one worked with the latest smartmontools: > > ID 04b4:6830 Cypress Semiconductor Corp. CY7C68300A EZ-USB AT2 USB 2.0 to ATA/ATAPI lsusb -v (or -vvv) Actually, the device provides the USB id. Identification is read from usb.ids file (its location depends on distribution). I am not sure how exactly Cypress chips are designed, but many chip types allow simple change of USB identification (using cheap external EEPROM). In this case, different USB IDs don't mean different hardware (and vice versa). -- Best Regards / S pozdravem, Stanislav Brabec software developer --------------------------------------------------------------------- SUSE LINUX, s. r. o. e-mail: sb...@su... Lihovarská 1060/12 tel: +420 284 028 966, +49 911 740538747 190 00 Praha 9 fax: +420 284 028 951 Czech Republic http://www.suse.cz/ |
From: Bruce A. <ba...@gr...> - 2008-06-09 15:42:21
|
>>> But you may check the usb.ids table. >>> >>> This one worked with the latest smartmontools: >>> ID 04b4:6830 Cypress Semiconductor Corp. CY7C68300A EZ-USB AT2 USB 2.0 to ATA/ATAPI > > lsusb -v (or -vvv) OK! > Actually, the device provides the USB id. > > Identification is read from usb.ids file (its location depends on > distribution). OK, I see. > I am not sure how exactly Cypress chips are designed, but many chip > types allow simple change of USB identification (using cheap external > EEPROM). In this case, different USB IDs don't mean different hardware > (and vice versa). Do I understand correctly: this means that some USB drive boxes with the Cypress chip might have a DIFFERENT USB ID than the one above? Cheers, Bruce |
From: Stanislav B. <sb...@su...> - 2008-06-09 16:51:56
|
Bruce Allen wrote: > > I am not sure how exactly Cypress chips are designed, but many chip > > types allow simple change of USB identification (using cheap external > > EEPROM). In this case, different USB IDs don't mean different hardware > > (and vice versa). > > Do I understand correctly: this means that some USB drive boxes with the > Cypress chip might have a DIFFERENT USB ID than the one above? It depends on chip design. If OEM branding allows change of USB ID, then yes. If OEM branding does not allow change of USB ID or it is done on contract base, Cypress Semiconductor Corp. should know a complete list of supported USB IDs. For example, this one seems to use Cypress chipset: 0547:2810 Anchor Chips, Inc. Cypress USB ATAPI Bridge http://www.sitecom.com/faq_result.php?productname=USB+2.0+IDE+adapter&productcode=CN-130&productid=298&subgroupid=20&groupid=8 For this one datasheet declares possibility to be used as ATA bridge, but it seems to be a generic chip, which does not implement any translation: 04b4:8613 Cypress Semiconductor Corp. CY7C68013 EZ-USB FX2 USB 2.0 Development Kit -- Best Regards / S pozdravem, Stanislav Brabec software developer --------------------------------------------------------------------- SUSE LINUX, s. r. o. e-mail: sb...@su... Lihovarská 1060/12 tel: +420 284 028 966, +49 911 740538747 190 00 Praha 9 fax: +420 284 028 951 Czech Republic http://www.suse.cz/ |
From: matthieu c. <cas...@fr...> - 2008-06-09 17:54:32
|
Stanislav Brabec wrote: > Bruce Allen wrote: > >>> I am not sure how exactly Cypress chips are designed, but many chip >>> types allow simple change of USB identification (using cheap external >>> EEPROM). In this case, different USB IDs don't mean different hardware >>> (and vice versa). >> Do I understand correctly: this means that some USB drive boxes with the >> Cypress chip might have a DIFFERENT USB ID than the one above? > > It depends on chip design. If OEM branding allows change of USB ID, then > yes. If OEM branding does not allow change of USB ID or it is done on > contract base, Cypress Semiconductor Corp. should know a complete list > of supported USB IDs. > To quote Cypress datasheet : Vendor ID. Cypress' Vendor ID may only be used for evalu- ation purposes, and not in released products. But it seems there is a lot's of manufacturer that reuse Cypress id... BTW it did a small test program for testing atacb. It can be download from http://castet.matthieu.free.fr/download/atacb-test.tar.gz > For example, this one seems to use Cypress chipset: > 0547:2810 Anchor Chips, Inc. Cypress USB ATAPI Bridge > http://www.sitecom.com/faq_result.php?productname=USB+2.0+IDE+adapter&productcode=CN-130&productid=298&subgroupid=20&groupid=8 IRRC Cypress brought Anchor Chips, and then with their technology did Cypress USB ATA Bridge chips. And this chip also support atacb because Linux kernel usb-storage implement ATA mode (the chip only support ATAPI mass storage) with atacb. Matthieu |
From: matthieu c. <cas...@fr...> - 2008-06-09 17:47:03
|
Stanislav Brabec wrote: > Douglas Gilbert wrote: >> >>> How can we tell if there is a Cypress chipset? >>> > > usb-storage has no reason to check for the chipset. Support for ATA > passthrough transport is above the scope of the driver. > For information, usb-storage will have a special mode for this chipset in 2.6.26. This will allow to use hdparm, hddtemp and other software that use SAT to works by doing a SAT <-> atacb translation. atacb support in smartmontools is still useful because it runs on other OS (Windows, solaris, ...), it can support atacb on older kernel version and it is more easier to test. Matthieu |
From: Bruce A. <ba...@gr...> - 2008-06-09 21:22:12
|
Matthieu, Doug, Something that would be useful would be to have smartmontools recognize a Cypress chipset device and return a useful hint (or automatically set the correct -d sat or -d usbcypress options). Is this possible? In this case smartctl -a /dev/sda would work if /dev/sda referred to an external USB disk with the correct Cypress chip. Or at least the user would get a clear error message saying "please add -d sat or -d usbcypress". Cheers, Bruce On Mon, 9 Jun 2008, matthieu castet wrote: > Stanislav Brabec wrote: >> Douglas Gilbert wrote: >>> >>>> How can we tell if there is a Cypress chipset? >>>> >> >> usb-storage has no reason to check for the chipset. Support for ATA >> passthrough transport is above the scope of the driver. >> > For information, usb-storage will have a special mode for this chipset > in 2.6.26. This will allow to use hdparm, hddtemp and other software > that use SAT to works by doing a SAT <-> atacb translation. > atacb support in smartmontools is still useful because it runs on other > OS (Windows, solaris, ...), it can support atacb on older kernel version > and it is more easier to test. > > Matthieu > > > ------------------------------------------------------------------------- > Check out the new SourceForge.net Marketplace. > It's the best place to buy or sell services for > just about anything Open Source. > http://sourceforge.net/services/buy/index.php > _______________________________________________ > Smartmontools-devel mailing list > Sma...@li... > https://lists.sourceforge.net/lists/listinfo/smartmontools-devel > |
From: matthieu c. <cas...@fr...> - 2008-06-09 21:36:30
|
Bruce Allen wrote: > Matthieu, Doug, > > Something that would be useful would be to have smartmontools recognize > a Cypress chipset device and return a useful hint (or automatically set > the correct -d sat or -d usbcypress options). Is this possible? I should be : try to send a command and see if it fails. My test program try ATA_OP_IDENTIFY and ATA_OP_PIDENTIFY. How does it works for sat ? Matthieu |
From: Bruce A. <ba...@gr...> - 2008-06-10 22:26:26
|
The first step is just to ask, what happens if a user does smartctl -a /dev/sda and /dev/sda is a USB Cypress device. Can we modify smartctl so that in this case it detects this case and prints a 'hint' for the user? Cheers, Bruce On Mon, 9 Jun 2008, matthieu castet wrote: > Bruce Allen wrote: >> Matthieu, Doug, >> >> Something that would be useful would be to have smartmontools recognize a >> Cypress chipset device and return a useful hint (or automatically set the >> correct -d sat or -d usbcypress options). Is this possible? > I should be : try to send a command and see if it fails. My test program try > ATA_OP_IDENTIFY and ATA_OP_PIDENTIFY. > > How does it works for sat ? > > Matthieu > |
From: Matthew D. <mdh...@on...> - 2008-06-10 22:42:07
|
The only way to know that it's a Cypress device is to look at the bridge ID and compare it to a list of known IDs. To get the bridge ID, the easiest way is to use sysfs to convert the /dev/sdX node into it's SCSI HBA and convert that into the USB device, then examine the ID information in the USB sysfs node. Matt On Tue, Jun 10, 2008 at 05:26:20PM -0500, Bruce Allen wrote: > The first step is just to ask, what happens if a user does > smartctl -a /dev/sda > and /dev/sda is a USB Cypress device. Can we modify smartctl so that in > this case it detects this case and prints a 'hint' for the user? > > Cheers, > Bruce > > > > On Mon, 9 Jun 2008, matthieu castet wrote: > > >Bruce Allen wrote: > >>Matthieu, Doug, > >> > >>Something that would be useful would be to have smartmontools recognize a > >>Cypress chipset device and return a useful hint (or automatically set the > >>correct -d sat or -d usbcypress options). Is this possible? > >I should be : try to send a command and see if it fails. My test program > >try ATA_OP_IDENTIFY and ATA_OP_PIDENTIFY. > > > >How does it works for sat ? > > > >Matthieu > > -- Matthew Dharm Home: mdh...@on... Maintainer, Linux USB Mass Storage Driver Umm, these aren't the droids you're looking for. -- Bill Gates User Friendly, 11/14/1998 |
From: Matthew D. <mdh...@on...> - 2008-06-04 04:39:31
|
I'm not entirely clear on the question, here... if I don't cover what you want to know, feel free to ask again. At the protocol level, usb-storage cares very little about the commands themselves. It basically packages SCSI commands (regardless of their source or content) for shipment across the wire to the device. That device may be a bridge or something else. That said, almost no USB devices support SMART commands. Even among ATA bridge chips, it's almost unheard of. There is support for the Cypress ATACB protocol (available on some of their bridges). That's fairly new, and will only work with certain Cypress bridges. If you have a kernel tree with these patches, you should be able to find a list of supported VID/PID codes in the unusual_devs.h file. As I recall, those required SCSI commands which were marked as "passthrough" commands somehow. The details are fuzzy (it's been a while since the patches were sent upstream) and I don't have a checked-out tree handy. Basically, this is a Cypress-proprietary way to package ATA commands over a SCSI-like interface. The one thing I want to make clear (which seems to be some confusion based on the e-mail below) is that the chipset in the computer doesn't matter at all. If it's UHCI, EHCI, or OHCI from Intel, Via, NEC, or whomever... these are just USB interface chips. What matters is the chip in the device at the other end of the USB cable. Matt On Tue, Jun 03, 2008 at 09:50:31PM -0400, Douglas Gilbert wrote: > Bruce Allen wrote: > >Could someone from the developers list respond to this? > > > >---------- Forwarded message ---------- > >Date: Sun, 25 May 2008 12:47:20 -0700 > >From: Jim Kleckner <jek...@kl...> > >To: sma...@li... > >Subject: [smartmontools-support] smartctl USB support (atacb) > > > >The recent patches for atacb support for USB drives > >with Cypress chipsets are exciting. > > > >Very often you are trying to deal with drives using a > >USB2.0 bridge like the one from Vantec. > > > >Can you describe how we can tell whether smartctl > >will work with some form of pass-through for a given > >hardware configuration? > > > >How can we tell if there is a Cypress chipset? > > > >A large number of motherboards use the Intel chipsets. > >Will there ever be a way to use those USB systems > >with smartctl? > > I went looking in Linux for any indication what the USB > storage class bridge device is and found nothing. I > doubt that the usb-storage driver in the Linux kernel > knows, but if it does, I can find no report. Perhaps > the driver maintainer may correct me. > > > Doug Gilbert > [cc-ed linux usb-storage driver maintainer] -- Matthew Dharm Home: mdh...@on... Maintainer, Linux USB Mass Storage Driver Why am I talking to a toilet brush? -- CEO User Friendly, 4/30/1998 |
From: Stanislav B. <sb...@su...> - 2008-06-09 14:33:20
|
Douglas Gilbert wrote: > > > How can we tell if there is a Cypress chipset? > > > > A large number of motherboards use the Intel chipsets. > > Will there ever be a way to use those USB systems > > with smartctl? > > I went looking in Linux for any indication what the USB > storage class bridge device is and found nothing. I > doubt that the usb-storage driver in the Linux kernel > knows, but if it does, I can find no report. Perhaps > the driver maintainer may correct me. usb-storage has no reason to check for the chipset. Support for ATA passthrough transport is above the scope of the driver. But you may check the usb.ids table. This one worked with the latest smartmontools: ID 04b4:6830 Cypress Semiconductor Corp. CY7C68300A EZ-USB AT2 USB 2.0 to ATA/ATAPI -- Best Regards / S pozdravem, Stanislav Brabec software developer --------------------------------------------------------------------- SUSE LINUX, s. r. o. e-mail: sb...@su... Lihovarská 1060/12 tel: +420 284 028 966, +49 911 740538747 190 00 Praha 9 fax: +420 284 028 951 Czech Republic http://www.suse.cz/ |