Re: [U-Boot] [PATCH] usb: fix usb start problem with SMSC USB hub and Toshiba USB stick
Hello,
i`m searching for a solution for a while, so i`m trying my luck here. I`m using Seagate Dockstar (ARM926EJ-S based) NAS device and installed Debian ARM on that as documented at http://jeff.doozan.com/debian/
While it`s running fine with a sandisk flash usb stick and a sd-card in a usb cardreader, i`m not able to get the thingie boot with 2 different real 2.5" ide disks attached to two different kinds of ide2usb converter cables.
The pogoplug linux (which runs on dockstar by default) detects the disks very fine and debian installation from there runs fine too, but booting from that always fails like this:
U-Boot 2010.06-00695-gbd23130-dirty (Aug 30 2010 - 23:04:56) Marvell-Dockstar/Pogoplug by Jeff Doozan Hit any key to stop autoboot: 0 (Re)start USB... USB: Register 10011 NbrPorts 1 USB EHCI 1.00 scanning bus for devices... 3 USB Device(s) found scanning bus for storage devices... 1 Storage Device(s) found ** Bad partition 1 ** ** Bad partition 1 ** Wrong Image Format for bootm command ERROR: can't get kernel image! stopping USB..
I found a LOT of references on the net where people report similar problems for sheevaplug, pogoplug, dockstar - even with flash based media and just came across this posting on u-boot list. looks related to me!?
i assume there is something which needs be fixed in u-boot and this is not storage-device specific but must be specific to the usb chipset or platform.
the problem is that i don`t know how to unbrick dockstar yet, but i`d like to let you know that a lot of people will be happy if this gets fixed. i will read deeper into how to build uboot and safely replace it on my device, if there is anything more i can do to provide input or do testing please let me know. i think i can dig out tons of reports of similar issues. as my "i replace my homeserver with a dockstar"-projects depends on this, maybe i can even donate a dockstar for the person willing or be able to fix this issue.
from another post:
Is this problem only valid for the EHCI code? I can imagine that the other host controllers (like UHCI and OHCI and so on) code suffer the same problem and need similar fixes... (At least I know that I have here a couple of USB sticks that show similar problems with OHCI ;-) )
If this helps - this is some info from a successful linux boot on that dockstar platform. usb runs perfectly when in linux, but it seems that there are problems only in u-boot.
[ 1.180000] ehci_marvell ehci_marvell.70059: Marvell Orion EHCI [ 1.190000] ehci_marvell ehci_marvell.70059: new USB bus registered, assigned bus number 1 [ 1.230000] ehci_marvell ehci_marvell.70059: irq 19, io base 0xf1050100 [ 1.250000] ehci_marvell ehci_marvell.70059: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004 [ 1.250000] usb usb1: configuration #1 chosen from 1 choice [ 1.260000] hub 1-0:1.0: USB hub found [ 1.260000] hub 1-0:1.0: 1 port detected [ 1.380000] ohci_hcd: 2006 August 04 USB 1.1 'Open' Host Controller (OHCI) Driver [ 1.380000] USB Universal Host Controller Interface driver v3.0 [ 1.660000] usb 1-1: new high speed USB device using ehci_marvell and address 2 [ 1.810000] usb 1-1: configuration #1 chosen from 1 choice [ 1.810000] hub 1-1:1.0: USB hub found [ 1.820000] hub 1-1:1.0: 4 ports detected [ 2.170000] usb 1-1.3: new high speed USB device using ehci_marvell and address 3 [ 2.270000] usb 1-1.3: configuration #1 chosen from 1 choice
regards roland
ps: i found the following line in common/usb_storage.c - so maybe we cannot expect the usb code to work with real disks at all !? /* * New Note: * Support for USB Mass Storage Devices (BBB) has been added. It has * only been tested with USB memory sticks. */
quote: [U-Boot] [PATCH] usb: fix usb start problem with SMSC USB hub and Toshiba USB stick Anatolij Gustschin agust at denx.de Tue Aug 31 20:45:14 CEST 2010 Previous message: [U-Boot] [PATCH] zlib: Add further watchdog reset calls Next message: [U-Boot] [PATCH] [v2] mpc8610: refactor DIU initialization code to use existing bitmap and logo features Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Checking the status field of the qTD token in the current code do not take into acount cases where endpoint stall (halted) bit is set together with some other status bits. As a result clearing stall on an endpoint won't be done if other status bits were set.
E.g. 'usb start' often fails with Toshiba USB stick 0x930/0x6545 connected to the SMSC USB 2.0 hub 0x424/0x2514. Debugging the issue showed that while bulk IN transfers with length of 13 or 18 the status field of the qTD token sometimes indicates trans- action error (XactErr) and sometimes additionally endpoint halted state. In the latter case resetting the USB device in error recovery code fails as no clear stall request on the endpoint will be done. The patch fixes status field checking code to properly handle endpoint halted state.
However this fix is not enough to solve 'usb start' problem with hub/stick combination mentioned above. Running with lot of debug code in ehci_submit_async() I've never seen the problem with usb stick recognition. After removing this debug code the similar problem sometimes showed up again. Therefore the patch also adds delay in ehci_submit_async() for above-mentioned hub/stick combination. Even without this delay the fix is an improvement since it fixes the problem with board freezy after subsequent failed 'usb start/stop' cycles as it was observed on mpc5121ads board.
Signed-off-by: Anatolij Gustschin <agust at denx.de> --- common/usb_storage.c | 5 +++-- drivers/usb/host/ehci-hcd.c | 8 ++++++++ 2 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/common/usb_storage.c b/common/usb_storage.c index 76949b8..5ca92c3 100644 --- a/common/usb_storage.c +++ b/common/usb_storage.c @@ -680,7 +680,8 @@ int usb_stor_BBB_transport(ccb *srb, struct us_data *us) result = usb_bulk_msg(us->pusb_dev, pipe, srb->pdata, srb->datalen, &data_actlen, USB_CNTL_TIMEOUT * 5); /* special handling of STALL in DATA phase */ - if ((result < 0) && (us->pusb_dev->status & USB_ST_STALLED)) { + if ((result < 0) && + (us->pusb_dev->status & (USB_ST_STALLED | USB_ST_CRC_ERR))) { USB_STOR_PRINTF("DATA:stall\n"); /* clear the STALL on the endpoint */ result = usb_stor_BBB_clear_endpt_stall(us, @@ -710,7 +711,7 @@ again:
/* special handling of STALL in STATUS phase */ if ((result < 0) && (retry < 1) && - (us->pusb_dev->status & USB_ST_STALLED)) { + (us->pusb_dev->status & (USB_ST_STALLED | USB_ST_CRC_ERR))) { USB_STOR_PRINTF("STATUS:stall\n"); /* clear the STALL on the endpoint */ result = usb_stor_BBB_clear_endpt_stall(us, us->ep_in); diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c index 37d056e..7463a75 100644 --- a/drivers/usb/host/ehci-hcd.c +++ b/drivers/usb/host/ehci-hcd.c @@ -430,6 +430,12 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer, usbsts = ehci_readl(&hcor->or_usbsts); ehci_writel(&hcor->or_usbsts, (usbsts & 0x3f));
+ if (dev->descriptor.idVendor == 0x930 && + dev->descriptor.idProduct == 0x6545 && + dev->parent->descriptor.idVendor == 0x424 && + dev->parent->descriptor.idProduct == 0x2514) + wait_ms(10); + /* Enable async. schedule. */ cmd = ehci_readl(&hcor->or_usbcmd); cmd |= CMD_ASE; @@ -490,6 +496,8 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer, break; default: dev->status = USB_ST_CRC_ERR; + if ((token & 0x40) == 0x40) + dev->status |= USB_ST_STALLED; break; } dev->act_len = length - ((token >> 16) & 0x7fff);
Dear "Roland Kletzing",
In message 9714EDEC3798462FACB91473A96284B8@samsungnetbook you wrote:
i`m searching for a solution for a while, so i`m trying my luck here. I`m using Seagate Dockstar (ARM926EJ-S based) NAS device and installed Debian ARM on that as documented at http://jeff.doozan.com/debian/
You hijack a (pretty much unrelated) mail thread.
You top-post.
You full-quote.
And it seems you did not read neither http://www.netmeister.org/news/learn2quote.html nor http://www.catb.org/~esr/faqs/smart-questions.html
I think your problem is completely unrelated to the Subject: you refer to.
Best regards,
Wolfgang Denk
pardon, that was not my intention, nor was it my intention to infringe list rules. so, this is for developers/experts - not for users and bug-reporters?
i`m spending more than half of this evening with this issue, trying to find a solution and trying to provide information for a problem affecting many people and which seems to exist in u-boot.
so please forgive if i jumped onto the wrong thread as an entry point and i helter-skelter blundered into this list. will go whine elsewhere then and use obscure workarounds for this issue instead ( like that guy on this thread: http://openplug.org/plugforum/index.php?topic=131.0 )
regards roland
----Ursprüngliche Nachricht----- Von: "Wolfgang Denk" wd@denx.de Gesendet: 20.09.2010 23:19:15 An: "Roland Kletzing" devzero@web.de Betreff: Re: [U-Boot] [PATCH] usb: fix usb start problem with SMSC USB hub and Toshiba USB stick
Dear "Roland Kletzing",
In message 9714EDEC3798462FACB91473A96284B8@samsungnetbook you wrote:
i`m searching for a solution for a while, so i`m trying my luck here. I`m using Seagate Dockstar (ARM926EJ-S based) NAS device and installed Debian ARM on that as documented at http://jeff.doozan.com/debian/
You hijack a (pretty much unrelated) mail thread.
You top-post.
You full-quote.
And it seems you did not read neither http://www.netmeister.org/news/learn2quote.html nor http://www.catb.org/~esr/faqs/smart-questions.html
I think your problem is completely unrelated to the Subject: you refer to.
Best regards,
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de I can't understand it. I can't even understand the people who can understand it. - Queen Juliana of the Netherlands.
___________________________________________________________ Neu: WEB.DE De-Mail - Einfach wie E-Mail, sicher wie ein Brief! Jetzt De-Mail-Adresse reservieren: https://produkte.web.de/go/demail02
Dear devzero@web.de,
In message 899533205.349626.1285022229052.JavaMail.fmail@mwmweb063 you wrote:
pardon, that was not my intention, nor was it my intention to infringe list rules. so, this is for developers/experts - not for users and bug-reporte rs?
This list is for ALL discussions about U-Boot: users, bug-reporters, developers, etc.
But we expect that everybody who posts here follows some basic Netiquette.
You don't. For example, Netiquette says to keep your line length below some 70 characters or so.
You don't even after being told. I asked you explicitly NOT to top post / full quote, and pointed you to rules for correct quoting. You ignore that.
i`m spending more than half of this evening with this issue, trying to find a solution and trying to provide information for a problem affecting manyr people and which seems to exist in u-boot.
I also pointed you to a document that explains how to ask smart questions, but it seems you ignored this hint as well.
You completely fail to explain which exact version of U-Boot you are running - it seems to be a locally modified one, probably with out-of-tree patches to support that Dockstar board which is not known in the mainline U-Boot repository. How should we be able to comment on code we don't even know?
Best regards,
Wolfgang Denk
Hi Roland,
i`m spending more than half of this evening with this issue, trying to find a solution and trying to provide information for a problem affecting many people and which seems to exist in u-boot.
Rest assured that many people on this list spend multiple days with USB problems :|
so please forgive if i jumped onto the wrong thread as an entry point and i helter-skelter blundered into this list. will go whine elsewhere then and use obscure workarounds for this issue instead ( like that guy on this thread: http://openplug.org/plugforum/index.php?topic=131.0 )
The guy doing a reboot when it fails? Actually not a bad strategy. It shows once more that USB has the tendency to be non-deterministic.
As always, it would be good to first find out if the codebase of the U-Boot you use is comparable to mainline (otherwise we cannot use _any_ tests you do) and then do systematic tests to narrow down the problem.
It may well be that USB has been tested better in U-Boot context with USB sticks (solid state) rather than mechainically spinning hard disks. Maybe this is something which needs to be taken into account in the code.
Keep us posted.
Cheers Detlev
Hello Detlev,
i`m spending more than half of this evening with this issue, trying to find a solution and trying to provide information for a problem affecting many people and which seems to exist in u-boot.
Rest assured that many people on this list spend multiple days with USB problems :|
Ok, we know that from linux..... :(
so please forgive if i jumped onto the wrong thread as an entry point and i helter-skelter blundered into this list. will go whine elsewhere then and use obscure workarounds for this issue instead ( like that guy on this thread: http://openplug.org/plugforum/index.php?topic=131.0 )
The guy doing a reboot when it fails? Actually not a bad strategy. It shows once more that USB has the tendency to be non-deterministic.
Yes, but for my setup unfortunately neither reboot (i.e. warm-boot) helps, nor does it help to add pre-boot usb reset to the u-boot config.
As always, it would be good to first find out if the codebase of the U-Boot you use is comparable to mainline (otherwise we cannot use _any_ tests you do) and then do systematic tests to narrow down the problem.
See my other posting.
It may well be that USB has been tested better in U-Boot context with USB sticks (solid state) rather than mechainically spinning hard disks. Maybe this is something which needs to be taken into account in the code.
That`s a good theory. btw, iirc, i saw some hint that the code (usb-storage.c !?) that it had only been well tested with flash media.
Anyway, just to tell about the background: I`m trying to replace my home server with a Dockstar. As this runs 24/7 i want something energy-efficient for that. I got some Dockstars for so cheap and they are asbolute fantastic and just perfect fit to be run with debian. I plan buying a solid state sata disk with sata2usb one day, but seeing two ordinary disks with 2 different ide2usb fail does not encourage me to spend bucks on something new. I want a robust system where i can use more than "some" drives which are just compatible by chance. From what i have tested, the debian install on dockstar runs pretty well, stable and fast.
regards Roland ___________________________________________________________ GRATIS: Spider-Man 1-3 sowie 300 weitere Videos! Jetzt kostenlose Movie-FLAT freischalten! http://movieflat.web.de
Hi Roland,
[...]
Anyway, just to tell about the background: I`m trying to replace my home server with a Dockstar. As this runs 24/7 i want something energy-efficient for that. I got some Dockstars for so cheap and they are asbolute fantastic and just perfect fit to be run with debian. I plan buying a solid state sata disk with sata2usb one day, but seeing two ordinary disks with 2 different ide2usb fail does not encourage me to spend bucks on something new. I want a robust system where i can use more than "some" drives which are just compatible by chance. From what i have tested, the debian install on dockstar runs pretty well, stable and fast.
In my personal experience, "robust" and "usb" do not mix well in one sentence. Maybe I'm somewhat conservative, but just in another product I have seen spurious USB dis- and reconnects after days and weeks of otherwise solid functioning. Of course they _did_ mess up the software and crash the system. So be sure that your software stack can cope with something like that.
In effect, currently I would not try to design a "robust system" (ready to be sold as a product) with its primary rootfilesystem on a USB storage device. For IDE/ATA/SATA all the problems in a running system that I have seen have had their cause in the device, not in the interconnect method. This is not something I can say about USB.
Cheers Detlev
Good evening !
Some new (real weird) findings i like to report.
I have tried a third usb2ide adapter cable of different brand. Had that around at work....
Apparently , it uses exactly the same jmicron chipset as one of the other ones which fail - but this one - surprise - WORKS ! The output of "usb info" is exactly the same. Where the former one fails with "usb part" , this one works with "usb part" and gives proper info and i can boot without a problem. Loading kernel & initrd is slow, but it works, even with cold boot.
So, i`m wondering who`s to blame - the cable or the sofware, or both :)
Since both cables work with linux, maybe there is some slight differenc in pcb/cable design and thus some slight timing difference which causing this problem.
Maybe i will go to find some more convertor cables, just for curiousity and to see how they behave. Does someone know a website listing usb convertor cables and compare their "quality" ? I did not find such....
Anyway, so my problem is not that serious anymore, as i know the hardware is part of the problem and which can be replaced.....
regards roland
Marvell>> usb info usb info 1: Hub, USB Revision 2.0 - u-boot EHCI Host Controller - Class: Hub - PacketSize: 64 Configurations: 1 - Vendor: 0x0000 Product 0x0000 Version 1.0 Configuration: 1 - Interfaces: 1 Self Powered 0mA Interface: 0 - Alternate Setting 0, Endpoints: 1 - Class Hub - Endpoint 1 In Interrupt MaxPacket 8 Interval 255ms
2: Hub, USB Revision 2.0 - USB2.0 Hub - Class: Hub - PacketSize: 64 Configurations: 1 - Vendor: 0x05e3 Product 0x0608 Version 6.3 Configuration: 1 - Interfaces: 1 Self Powered Remote Wakeup 100mA Interface: 0 - Alternate Setting 0, Endpoints: 1 - Class Hub - Endpoint 1 In Interrupt MaxPacket 1 Interval 12ms
3: Mass Storage, USB Revision 2.0 - JMicron USB to ATA/ATAPI Bridge 222222126873 - Class: (from Interface) Mass Storage - PacketSize: 64 Configurations: 1 - Vendor: 0x152d Product 0x2338 Version 1.0 Configuration: 1 - Interfaces: 1 Self Powered 2mA - String: "USB Mass Storage" Interface: 0 - Alternate Setting 0, Endpoints: 2 - Class Mass Storage, Transp. SCSI, Bulk only - String: "MSC Bulk-Only Transfer" - Endpoint 1 In Bulk MaxPacket 512 - Endpoint 2 Out Bulk MaxPacket 512
Marvell>> usb storage usb storage Device 0: Vendor: HITACHI_ Rev: A0A5 Prod: DK14FA-20 Type: Hard Disk Capacity: 19077.1 MB = 18.6 GB (39070080 x 512)
usb part
Partition Map for USB device 0 -- Partition Type: DOS
Partition Start Sector Num Sectors Type 1 63 4016187 83 2 4016250 514080 82
-------------------------------------------------- From: "Detlev Zundel" dzu@denx.de Sent: Thursday, September 23, 2010 11:24 AM To: devzero@web.de Cc: "Wolfgang Denk" wd@denx.de; u-boot@lists.denx.de Subject: Re: [U-Boot] [PATCH] usb: fix usb start problem with SMSC USB hub and Toshiba USB stick
Hi Roland,
[...]
Anyway, just to tell about the background: I`m trying to replace my home server with a Dockstar. As this runs 24/7 i want something energy-efficient for that. I got some Dockstars for so cheap and they are asbolute fantastic and just perfect fit to be run with debian. I plan buying a solid state sata disk with sata2usb one day, but seeing two ordinary disks with 2 different ide2usb fail does not encourage me to spend bucks on something new. I want a robust system where i can use more than "some" drives which are just compatible by chance. From what i have tested, the debian install on dockstar runs pretty well, stable and fast.
In my personal experience, "robust" and "usb" do not mix well in one sentence. Maybe I'm somewhat conservative, but just in another product I have seen spurious USB dis- and reconnects after days and weeks of otherwise solid functioning. Of course they _did_ mess up the software and crash the system. So be sure that your software stack can cope with something like that.
In effect, currently I would not try to design a "robust system" (ready to be sold as a product) with its primary rootfilesystem on a USB storage device. For IDE/ATA/SATA all the problems in a running system that I have seen have had their cause in the device, not in the interconnect method. This is not something I can say about USB.
Cheers Detlev
-- error compiling committee.c: too many arguments to function -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu@denx.de
Hi Roland,
[...]
I have tried a third usb2ide adapter cable of different brand. Had that around at work....
Apparently , it uses exactly the same jmicron chipset as one of the other ones which fail - but this one - surprise - WORKS !
Just as an additional datapoint - we have also seen usb sticks which by all available markings looked to contain the exact same silicon. Only persistent questioning of the manufacturer in the end showed that "harmless changes" were done and were deemed not important enough to be marked by changing serial-numbers or other visible identification.
The output of "usb info" is exactly the same. Where the former one fails with "usb part" , this one works with "usb part" and gives proper info and i can boot without a problem. Loading kernel & initrd is slow, but it works, even with cold boot.
So, i`m wondering who`s to blame - the cable or the sofware, or both :)
I still blame the inventors of USB. One of them is even depicted as a rock star in an Intel ad - ugh. I wonder what a short film with some actual USB developers would look like :)
Since both cables work with linux, maybe there is some slight differenc in pcb/cable design and thus some slight timing difference which causing this problem.
Maybe i will go to find some more convertor cables, just for curiousity and to see how they behave. Does someone know a website listing usb convertor cables and compare their "quality" ? I did not find such....
Did you try http://www.qbik.ch/usb/devices/ ?
Anyway, so my problem is not that serious anymore, as i know the hardware is part of the problem and which can be replaced.....
See, that's another reason why we still have such problems - once a hardware replacement is found, the problem is no longer pursued.
Note that this is a not meant as a criticism, but simply as an explanation of the current state of affairs - I've seen this multiple times now.
Cheers Detlev
[1] http://www.shoppingblog.com/blog/517099
hi,
Maybe i will go to find some more convertor cables, just for curiousity and to see how they behave. Does someone know a website listing usb convertor cables and compare their "quality" ? I did not find such....
Did you try http://www.qbik.ch/usb/devices/ ?
great link, thanks !
See, that's another reason why we still have such problems - once a hardware replacement is found, the problem is no longer pursued.
Oh, have a look on the linux bugtracker or other linux communities. I have seen so many problems go away by hardware replacement ("sorry, i can`t reproduce the problem, i gave back the mobo and bought a new one...."). But even more problems were not solved because either nobody cared or nobody cared soon enough.....
Anyway, i found one of the problematic USB bridge cables being problematic on my new netbook, too (disk not being recognized), so i assume it`s really crap and my 2 failing setups are mostly due to bad hardware. I will keep an eye on that and report my findings, if appreciated.
regards roland
Hi Roland,
[...]
See, that's another reason why we still have such problems - once a hardware replacement is found, the problem is no longer pursued.
Oh, have a look on the linux bugtracker or other linux communities.
USB support in Linux is way better than what we have in U-Boot and thats what I was talking about.
Anyway, i found one of the problematic USB bridge cables being problematic on my new netbook, too (disk not being recognized), so i assume it`s really crap and my 2 failing setups are mostly due to bad hardware. I will keep an eye on that and report my findings, if appreciated.
Of course this is appreciated! Many samples in the "experince space" are important information for developers. Although we usually only see the "it doesn't work" samples, some "hey it works!" would also not hurt ;)
Cheers Detlev
participants (4)
-
Detlev Zundel -
devzero＠web.de -
Roland Kletzing
-
Wolfgang Denk