Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Alan Stern wrote: I just tried doing what you said: modprobe net2280 dd if=/dev/zero of=/tmp/b bs=1M count=5 ; \ modprobe g_file_storage file=/tmp/b No problem; it worked fine. But my version of the driver includes the patch, of course. And this was under 2.6.22-rc3. Right, I did the same (ok, 4M) and it worked too. Then I did it with my usual size - 16M - and it broke. Also with the patch you suggested. Thanks Guennadi - Guennadi Liakhovetski, Ph.D. DSA Daten- und Systemtechnik GmbH Pascalstr. 28 D-52076 Aachen Germany - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Wed, 6 Jun 2007, Guennadi Liakhovetski wrote: On Tue, 5 Jun 2007, Alan Stern wrote: I just tried doing what you said: modprobe net2280 dd if=/dev/zero of=/tmp/b bs=1M count=5 ; \ modprobe g_file_storage file=/tmp/b No problem; it worked fine. But my version of the driver includes the patch, of course. And this was under 2.6.22-rc3. Right, I did the same (ok, 4M) and it worked too. Then I did it with my usual size - 16M - and it broke. Also with the patch you suggested. Okay. I just tried it over again, this time using 16 MB. Just as in your case, the filesystem was standard ext3. And it worked. So you'll have to do some debugging on your own. The source file is drivers/usb/gadget/file_storage.c. To begin with, you can #define the DEBUG and VERBOSE macros near the start of the file. Some other things to check: Make sure that fsg_bind() completes successfully. See if handle_exception() is getting called and if it is calling do_set_config() under the FSG_STATE_CONFIG_CHANGE case. Also see if do_set_config() returns an error. Alan Stern - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Guennadi Liakhovetski wrote: Hi I see a ~100% reproducible problem under 2.6.20 with g_file_storage: the script cp fs.image /tmp/ modprobe g_file_storage file=/tmp/fs.image comes with no error back, but the enumeration fails. Whereas inserting a sync between the cp and modprobe fixes it... It worked without sync under 2.4.30. Filesystem is ext3 mounted with default parameters (rw,data=ordered), gadget driver net2280, i386 platform. Is this intended behaviour? Yes, of course we always write device drivers so that they will fail in unexpected ways unless the user does something unintuitive and undocumented. :-) When you say the enumeration fails, what exactly goes wrong? I've never tried using g_file_storage with a newly-created backing file. Alan Stern - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
Alan Stern wrote: Yes, of course we always write device drivers so that they will fail in unexpected ways unless the user does something unintuitive and undocumented. :-) Ya, we all knew that. Just so long as you understand that we users all read the hardware docs from cover to cover, looking for all those undocumented features just so we can justify the errant driver report. :} - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Alan Stern wrote: On Tue, 5 Jun 2007, Guennadi Liakhovetski wrote: I see a ~100% reproducible problem under 2.6.20 with g_file_storage: the script cp fs.image /tmp/ modprobe g_file_storage file=/tmp/fs.image comes with no error back, but the enumeration fails. Whereas inserting a sync between the cp and modprobe fixes it... It worked without sync under 2.4.30. Filesystem is ext3 mounted with default parameters (rw,data=ordered), gadget driver net2280, i386 platform. Is this intended behaviour? Yes, of course we always write device drivers so that they will fail in unexpected ways unless the user does something unintuitive and undocumented. :-) Sure, I always program that way too:-) When you say the enumeration fails, what exactly goes wrong? I've never tried using g_file_storage with a newly-created backing file. Ok, the host says: usb usb1: wakeup_rh (auto-start) hub 1-0:1.0: state 7 ports 2 chg evt 0004 uhci_hcd :00:1f.2: port 2 portsc 0093,00 hub 1-0:1.0: port 2, status 0101, change 0001, 12 Mb/s hub 1-0:1.0: debounce: port 2: total 100ms stable 100ms status 0x101 usb 1-2: new full speed USB device using uhci_hcd and address 69 usb 1-2: default language 0x0409 usb 1-2: new device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-2: Product: File-backed Storage Gadget usb 1-2: Manufacturer: Linux 2.6.20.1-rt8-dsa-mftd2x-2 with net2280 usb 1-2: SerialNumber: 3238204E6F76 usb 1-2: uevent usb 1-2: configuration #1 chosen from 1 choice usb 1-2: khubd timed out on ep0out len=0/0 usb 1-2: can't set config #1, error -110 drivers/usb/core/inode.c: creating file '069' hub 1-0:1.0: state 7 ports 2 chg evt 0004 and on USB-analyser I see an invalid out-transaction with an invalid toggle and a setup transactio with only NAKs. Thanks Guennadi - Guennadi Liakhovetski, Ph.D. DSA Daten- und Systemtechnik GmbH Pascalstr. 28 D-52076 Aachen Germany - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Guennadi Liakhovetski wrote: When you say the enumeration fails, what exactly goes wrong? I've never tried using g_file_storage with a newly-created backing file. Ok, the host says: usb usb1: wakeup_rh (auto-start) hub 1-0:1.0: state 7 ports 2 chg evt 0004 uhci_hcd :00:1f.2: port 2 portsc 0093,00 hub 1-0:1.0: port 2, status 0101, change 0001, 12 Mb/s hub 1-0:1.0: debounce: port 2: total 100ms stable 100ms status 0x101 usb 1-2: new full speed USB device using uhci_hcd and address 69 usb 1-2: default language 0x0409 usb 1-2: new device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-2: Product: File-backed Storage Gadget usb 1-2: Manufacturer: Linux 2.6.20.1-rt8-dsa-mftd2x-2 with net2280 usb 1-2: SerialNumber: 3238204E6F76 usb 1-2: uevent usb 1-2: configuration #1 chosen from 1 choice usb 1-2: khubd timed out on ep0out len=0/0 usb 1-2: can't set config #1, error -110 drivers/usb/core/inode.c: creating file '069' hub 1-0:1.0: state 7 ports 2 chg evt 0004 and on USB-analyser I see an invalid out-transaction with an invalid toggle and a setup transactio with only NAKs. This looks like it has nothing to do with using sync or newly-created backing files. Instead it resembles a problem for which a patch was recently submitted: http://marc.info/?l=linux-usb-develm=118053809808501w=2 It hasn't been accepted yet, unfortunately. If you can confirm that it fixes your problem, I'll urge Greg to merge it in time for 2.6.22. Alan Stern - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Alan Stern wrote: On Tue, 5 Jun 2007, Guennadi Liakhovetski wrote: usb usb1: wakeup_rh (auto-start) hub 1-0:1.0: state 7 ports 2 chg evt 0004 uhci_hcd :00:1f.2: port 2 portsc 0093,00 hub 1-0:1.0: port 2, status 0101, change 0001, 12 Mb/s hub 1-0:1.0: debounce: port 2: total 100ms stable 100ms status 0x101 usb 1-2: new full speed USB device using uhci_hcd and address 69 usb 1-2: default language 0x0409 usb 1-2: new device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-2: Product: File-backed Storage Gadget usb 1-2: Manufacturer: Linux 2.6.20.1-rt8-dsa-mftd2x-2 with net2280 usb 1-2: SerialNumber: 3238204E6F76 usb 1-2: uevent usb 1-2: configuration #1 chosen from 1 choice usb 1-2: khubd timed out on ep0out len=0/0 usb 1-2: can't set config #1, error -110 drivers/usb/core/inode.c: creating file '069' hub 1-0:1.0: state 7 ports 2 chg evt 0004 and on USB-analyser I see an invalid out-transaction with an invalid toggle and a setup transactio with only NAKs. This looks like it has nothing to do with using sync or newly-created backing files. Instead it resembles a problem for which a patch was recently submitted: http://marc.info/?l=linux-usb-develm=118053809808501w=2 Emn, how recent are those signal-handling changes? Remember, the kernel on the gadget is only 2.6.20.1. And macroscopically it is indeed 100% repropducible - with and without sync. Can it really be explained by this signal-handling problem? It hasn't been accepted yet, unfortunately. If you can confirm that it fixes your problem, I'll urge Greg to merge it in time for 2.6.22. I'll try to test it tomorrow. Thanks Guennadi - Guennadi Liakhovetski, Ph.D. DSA Daten- und Systemtechnik GmbH Pascalstr. 28 D-52076 Aachen Germany - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
Re: [linux-usb-devel] file_storage gadget requires a sync...
On Tue, 5 Jun 2007, Guennadi Liakhovetski wrote: This looks like it has nothing to do with using sync or newly-created backing files. Instead it resembles a problem for which a patch was recently submitted: http://marc.info/?l=linux-usb-develm=118053809808501w=2 Emn, how recent are those signal-handling changes? Remember, the kernel on the gadget is only 2.6.20.1. And macroscopically it is indeed 100% repropducible - with and without sync. Can it really be explained by this signal-handling problem? I don't know; I didn't follow the changes to the core signal-handling for kernel threads. I never experienced any problems until I started using 2.6.22, so it's certainly possible that you're seeing something different. I just tried doing what you said: modprobe net2280 dd if=/dev/zero of=/tmp/b bs=1M count=5 ; \ modprobe g_file_storage file=/tmp/b No problem; it worked fine. But my version of the driver includes the patch, of course. And this was under 2.6.22-rc3. Alan Stern - This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ ___ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
