Hi,
can anyone confirm that writing to usb-storage devices is working on SMP
systems? Especially to a SD Card in an USB 1.1 card reader attached to a
64-bit Dual Opteron with NUMA enabled?
I have a noname USB 1.1 card reader that identifies itself as 0dda:0001
'Integrated Circuit Solution, Inc.' attached to a Tyan Thunder K8W
(S2885) dual Opteron NUMA system. After copying data to the disk, the
activity light flashes for some time and the following messages appear
in syslog:
2005-01-15 13:41:39 +0100 kernel: ohci_hcd 0000:03:00.1: urb 0000010068400d40 path 1 ep2out 6fce0000 cc 6 --> status -71
2005-01-15 13:42:59 +0100 kernel: ohci_hcd 0000:03:00.1: GetStatus roothub.portstatus [0] = 0x00100103 PRSC PPS PES CCS
2005-01-15 13:42:59 +0100 kernel: usb 2-1: reset full speed USB device using ohci_hcd and address 10
2005-01-15 13:42:59 +0100 kernel: ohci_hcd 0000:03:00.1: GetStatus roothub.portstatus [0] = 0x00100103 PRSC PPS PES CCS
2005-01-15 13:42:59 +0100 kernel: usb 2-1: ep0 maxpacket = 8
2005-01-15 13:42:59 +0100 kernel: usb 2-1: manual set_interface for iface 0, alt 0
2005-01-15 13:43:15 +0100 kernel: scsi: Device offlined - not ready after error recovery: host 7 channel 0 id 0 lun 2
2005-01-15 13:43:15 +0100 kernel: SCSI error : <7 0 0 2> return code = 0x50000
2005-01-15 13:43:15 +0100 kernel: end_request: I/O error, dev sdc, sector 147093
2005-01-15 13:43:15 +0100 kernel: printk: 310019 messages suppressed.
2005-01-15 13:43:15 +0100 kernel: Buffer I/O error on device sdc, logical block 147093
The system behaves as expected if I boot the kernel with 'maxcpus=1'.
Tom
--
T h o m a s Z e h e t b a u e r ( TZ251 )
PGP encrypted mail preferred - KeyID 96FFCB89
finger [email protected] for key
Experience is what you get when you expected something else.
On Monday, 17 of January 2005 18:17, Thomas Zehetbauer wrote:
> Hi,
>
> can anyone confirm that writing to usb-storage devices is working on SMP
> systems?
Generally, it is. Recently, I've written some stuff to a USB pendrive (using
2.6.10-ac7 or -ac9).
Greets,
RJW
--
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
-- Lewis Carroll "Alice's Adventures in Wonderland"
On Llu, 2005-01-17 at 17:26, Rafael J. Wysocki wrote:
> On Monday, 17 of January 2005 18:17, Thomas Zehetbauer wrote:
> > Hi,
> >
> > can anyone confirm that writing to usb-storage devices is working on SMP
> > systems?
>
> Generally, it is. Recently, I've written some stuff to a USB pendrive (using
> 2.6.10-ac7 or -ac9).
I'm dumping stuff to large hard disks via USB2 with 2.6.9 and 2.6.10
reliably. With 2.4.x until very late and with 2.6.early it would hang
reliably. It now seems pretty solid. It does panic the box if you remove
the drive while its in use which IMHO is *bad* but thats harder to fix.
On Mon, Jan 17, 2005 at 06:26:33PM +0100, Rafael J. Wysocki wrote:
> On Monday, 17 of January 2005 18:17, Thomas Zehetbauer wrote:
> > Hi,
> >
> > can anyone confirm that writing to usb-storage devices is working on SMP
> > systems?
>
> Generally, it is. Recently, I've written some stuff to a USB pendrive (using
> 2.6.10-ac7 or -ac9).
Same here with Abit VP6 dual-P3 and 2.6.10. It shows up as /dev/sda,
and I can do anything that I would do with normal harddisk.
But, I still can't boot from it. :/ I can now mount it as root
filesystem, but I can't load the kernel from USB key drive.
--
William Park <[email protected]>, Toronto, Canada
Slackware Linux -- because I can type.
William Park wrote:
> On Mon, Jan 17, 2005 at 06:26:33PM +0100, Rafael J. Wysocki wrote:
>
>>On Monday, 17 of January 2005 18:17, Thomas Zehetbauer wrote:
>>
>>>Hi,
>>>
>>>can anyone confirm that writing to usb-storage devices is working on SMP
>>>systems?
>>
>>Generally, it is. Recently, I've written some stuff to a USB pendrive (using
>>2.6.10-ac7 or -ac9).
>
>
> Same here with Abit VP6 dual-P3 and 2.6.10. It shows up as /dev/sda,
> and I can do anything that I would do with normal harddisk.
>
> But, I still can't boot from it. :/ I can now mount it as root
> filesystem, but I can't load the kernel from USB key drive.
huh?? Who's mounting the root filesystem, then :) ?
If you mean that you can't get the BIOS to load the kernel for you, and
you're loading the kernel from a floppy or something, you should know
that some BIOS are pretty selective about what they consider a valid
boot partition.
I recommend that you use fdisk to set up one partition as FAT16 type
(even if you use another filesystem later), and make the partition
active. You might need to write a proper MBR on the pen also (IIRC LILO
as an option to do this).
You might also need to pass a special "disk=/dev/sda bios=0x80" (or
something like that) option in your lilo.conf file, but that depends how
far in the boot process you're hanging.
--
Paulo Marques - http://www.grupopie.com
"A journey of a thousand miles begins with a single step."
Lao-tzu, The Way of Lao-tzu
On Mon, Jan 17, 2005 at 08:01:23PM +0000, Paulo Marques wrote:
> William Park wrote:
> >On Mon, Jan 17, 2005 at 06:26:33PM +0100, Rafael J. Wysocki wrote:
> >
> >>On Monday, 17 of January 2005 18:17, Thomas Zehetbauer wrote:
> >>
> >>>Hi,
> >>>
> >>>can anyone confirm that writing to usb-storage devices is working
> >>>on SMP systems?
> >>
> >>Generally, it is. Recently, I've written some stuff to a USB
> >>pendrive (using 2.6.10-ac7 or -ac9).
> >
> >
> >Same here with Abit VP6 dual-P3 and 2.6.10. It shows up as /dev/sda,
> >and I can do anything that I would do with normal harddisk.
> >
> >But, I still can't boot from it. :/ I can now mount it as root
> >filesystem, but I can't load the kernel from USB key drive.
>
> huh?? Who's mounting the root filesystem, then :) ?
>
> If you mean that you can't get the BIOS to load the kernel for you, and
> you're loading the kernel from a floppy or something, you should know
> that some BIOS are pretty selective about what they consider a valid
> boot partition.
Hee, hee... yes, for now I boot from floppy first, then mount /dev/sda1
as root filesystem. I, of course, want to get rid of the floppy. My
motherboard is Mercury PVCLE266M-L with VIA C3 Samuel 2 cpu integrated
on the board. It has tons of boot device options, and detects my USB
key (SanDisk Cruzer Mini 256MB) as "USB RMD-FDD" and not as some kind of
harddisk. I think that's the problem: motherboard thinks it's floppy,
and LILO thinks it's SCSI harddisk (/dev/sda).
>
> I recommend that you use fdisk to set up one partition as FAT16 type
> (even if you use another filesystem later), and make the partition
> active. You might need to write a proper MBR on the pen also (IIRC
> LILO as an option to do this).
>
> You might also need to pass a special "disk=/dev/sda bios=0x80" (or
> something like that) option in your lilo.conf file, but that depends
> how far in the boot process you're hanging.
--
William Park <[email protected]>, Toronto, Canada
Slackware Linux -- because I can type.
Now you are at the topic...
I've had trouble mounting USB units ever since 2.6.7. What changed in the USB
Mass Storage department? :)
The log says:
Jan 17 23:41:24 qbox usb 5-2: new full speed USB device using uhci_hcd and
address 2
Jan 17 23:41:24 qbox usb-storage: USB Mass Storage device detected
Jan 17 23:41:24 qbox usb-storage: -- associate_dev
Jan 17 23:41:24 qbox usb-storage: Vendor: 0x07b4, Product: 0x0102, Revision:
0x1011
Jan 17 23:41:24 qbox usb-storage: Interface Subclass: 0x06, Protocol: 0x50
Jan 17 23:41:24 qbox usb-storage: Vendor: OLYMPUS, Product: u10D,S300D,u300D
Jan 17 23:41:24 qbox usb-storage: Transport: Bulk
Jan 17 23:41:24 qbox usb-storage: Protocol: Transparent SCSI
Jan 17 23:41:24 qbox usb-storage: usb_stor_control_msg: rq=fe rqtype=a1
value=0000 index=00 len=1
Jan 17 23:41:24 qbox usb-storage: GetMaxLUN command result is 1, data is 0
Jan 17 23:41:24 qbox usb-storage: *** thread sleeping.
Jan 17 23:41:24 qbox scsi0 : SCSI emulation for USB Mass Storage devices
Jan 17 23:41:24 qbox usb-storage: device found at 2
Jan 17 23:41:24 qbox usb-storage: waiting for device to settle before scanning
So that is okay, but when I try to mount it:
mount -t vfat /dev/scsi/host0/bus0/target0/lun0/part1 /tmp/CAM
mount: wrong fs type, bad option, bad superblock
on /dev/scsi/host0/bus0/target0/lun0/part1,
or too many mounted file systems
Does anyone by chance have a solution to this? (2.6.10 .config attached)
I'll be happy to buy a beer for the one fixing this! :-)
Best regards,
Kristian S?rensen.
On Monday 17 January 2005 18:17, Thomas Zehetbauer wrote:
> Hi,
>
> can anyone confirm that writing to usb-storage devices is working on SMP
> systems? Especially to a SD Card in an USB 1.1 card reader attached to a
> 64-bit Dual Opteron with NUMA enabled?
>
> I have a noname USB 1.1 card reader that identifies itself as 0dda:0001
> 'Integrated Circuit Solution, Inc.' attached to a Tyan Thunder K8W
> (S2885) dual Opteron NUMA system. After copying data to the disk, the
> activity light flashes for some time and the following messages appear
> in syslog:
>
> 2005-01-15 13:41:39 +0100 kernel: ohci_hcd 0000:03:00.1: urb
> 0000010068400d40 path 1 ep2out 6fce0000 cc 6 --> status -71 2005-01-15
> 13:42:59 +0100 kernel: ohci_hcd 0000:03:00.1: GetStatus roothub.portstatus
> [0] = 0x00100103 PRSC PPS PES CCS 2005-01-15 13:42:59 +0100 kernel: usb
> 2-1: reset full speed USB device using ohci_hcd and address 10 2005-01-15
> 13:42:59 +0100 kernel: ohci_hcd 0000:03:00.1: GetStatus roothub.portstatus
> [0] = 0x00100103 PRSC PPS PES CCS 2005-01-15 13:42:59 +0100 kernel: usb
> 2-1: ep0 maxpacket = 8
> 2005-01-15 13:42:59 +0100 kernel: usb 2-1: manual set_interface for iface
> 0, alt 0 2005-01-15 13:43:15 +0100 kernel: scsi: Device offlined - not
> ready after error recovery: host 7 channel 0 id 0 lun 2 2005-01-15 13:43:15
> +0100 kernel: SCSI error : <7 0 0 2> return code = 0x50000 2005-01-15
> 13:43:15 +0100 kernel: end_request: I/O error, dev sdc, sector 147093
> 2005-01-15 13:43:15 +0100 kernel: printk: 310019 messages suppressed.
> 2005-01-15 13:43:15 +0100 kernel: Buffer I/O error on device sdc, logical
> block 147093
>
> The system behaves as expected if I boot the kernel with 'maxcpus=1'.
>
> Tom
--
Kristian S?rensen
- The Umbrella Project -- Security for Consumer Electronics
http://umbrella.sourceforge.net
E-mail: [email protected], Phone: +45 29723816
I forgot to mention that I am using a 2.6.10 kernel but will try
2.6.11rc1 soon. Are you using a SMP system? I assume the cabling, card
reader and SD card are OK because the problem went away with maxcpus=1,
the problem must be in the USB ehci/ohci/storage drivers. Do you know if
there is any difference between addressing external hard disks and SD
cards?
Tom
--
T h o m a s Z e h e t b a u e r ( TZ251 )
PGP encrypted mail preferred - KeyID 96FFCB89
finger [email protected] for key
Microsoft Windows(tm). A thirty-two bit extension and graphical shell
to a sixteen bit patch to an eight bit operating system originally
coded for a four bit microprocessor which was written by a two-bit
company that can't stand one bit of competition.
It seems now that the problem was caused by my 6in1 Card Reader that
identifies itself as "ID 0dda:0001 Integrated Circuit Solution, Inc.".
Strange thing is that the problem went away with maxcpus=1 for my 256MB
SD-Cards but not with my new 1GB SD-Card now.
Tom
--
T h o m a s Z e h e t b a u e r ( TZ251 )
PGP encrypted mail preferred - KeyID 96FFCB89
finger [email protected] for key
Microsoft Windows(tm). A thirty-two bit extension and graphical shell
to a sixteen bit patch to an eight bit operating system originally
coded for a four bit microprocessor which was written by a two-bit
company that can't stand one bit of competition.