[gentoo-user] USB stick recognition problem

[gentoo-user] USB stick recognition problem

[Helmut Jarausch]

Hi,

since a few weeks I have a strange effect with my USB stick.

According to fdisk there is one partition on it
/dev/sde1              38     7839719     3919841    b  W95 FAT32

which I haven't changed for a long time.

Whenever I insert this stick, the kernel log shows
/dev/sde  but not /dev/sde1  (and there is no file /dev/sde1)

After Invoking fdisk /dev/sde with a simple 'p' command but nothing 
else, this device shows up.

Has anybody an idea what's going on here?

Many thanks for a hint,
Helmut.

Re: [gentoo-user] USB stick recognition problem

[Gregory SACRE]

Hi Helmut,


It sounds like there is a problem with the partition table on your USB
stick. It might be the consequence of a hardware failure (read or more
probably write) at some point.
If, as you mention, once you do an fdisk then "p", you can use once
again your USB stick, then maybe save everything you have on it, then
create a new partition table using fdisk on that disk and format the
whole thing.


Hope that helps,

Greg

On Mon, Jan 31, 2011 at 10:43 AM, Helmut Jarausch
<jarausch@igpm.rwth-aachen.de> wrote:
> Hi,
>
> since a few weeks I have a strange effect with my USB stick.
>
> According to fdisk there is one partition on it
> /dev/sde1              38     7839719     3919841    b  W95 FAT32
>
> which I haven't changed for a long time.
>
> Whenever I insert this stick, the kernel log shows
> /dev/sde  but not /dev/sde1  (and there is no file /dev/sde1)
>
> After Invoking fdisk /dev/sde with a simple 'p' command but nothing
> else, this device shows up.
>
> Has anybody an idea what's going on here?
>
> Many thanks for a hint,
> Helmut.
>
>

Re: [gentoo-user] USB stick recognition problem

[Paul Hartman]

On Mon, Jan 31, 2011 at 3:43 AM, Helmut Jarausch
<jarausch@igpm.rwth-aachen.de> wrote:
> Hi,
>
> since a few weeks I have a strange effect with my USB stick.
>
> According to fdisk there is one partition on it
> /dev/sde1              38     7839719     3919841    b  W95 FAT32
>
> which I haven't changed for a long time.
>
> Whenever I insert this stick, the kernel log shows
> /dev/sde  but not /dev/sde1  (and there is no file /dev/sde1)
>
> After Invoking fdisk /dev/sde with a simple 'p' command but nothing
> else, this device shows up.
>
> Has anybody an idea what's going on here?

I have a USB SD-card reader which cannot read the partition table when
first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
the kernel re-reads the partition table and everything works fine
after that. You can try it, maybe it'll work for you, too.

[gentoo-user] Re: USB stick recognition problem

[walt]

On 02/02/2011 07:48 AM, Paul Hartman wrote:

> I have a USB SD-card reader which cannot read the partition table when
> first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
> the kernel re-reads the partition table and everything works fine
> after that.

That sounds to me like a bug :)  Is it 100% reproducible, and do you see
the same  on other computers?

And, let us not forget the Ultimate Gold Standard: does it work on Windows?

Re: [gentoo-user] Re: USB stick recognition problem

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 662 bytes --]

On Wednesday 02 February 2011 21:29:22 walt wrote:
> On 02/02/2011 07:48 AM, Paul Hartman wrote:
> > I have a USB SD-card reader which cannot read the partition table when
> > first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
> > the kernel re-reads the partition table and everything works fine
> > after that.
> 
> That sounds to me like a bug :)  Is it 100% reproducible, and do you see
> the same  on other computers?
> 
> And, let us not forget the Ultimate Gold Standard: does it work on Windows?

Ha!  I usually invert this test when people tell me their MSWindows machine is 
playing up (again ...)
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] Re: USB stick recognition problem

[Paul Hartman]

On Wed, Feb 2, 2011 at 3:29 PM, walt <w41ter@gmail.com> wrote:
> On 02/02/2011 07:48 AM, Paul Hartman wrote:
>
>> I have a USB SD-card reader which cannot read the partition table when
>> first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
>> the kernel re-reads the partition table and everything works fine
>> after that.
>
> That sounds to me like a bug :)  Is it 100% reproducible, and do you see
> the same  on other computers?
>
> And, let us not forget the Ultimate Gold Standard: does it work on Windows?

Yes, it is exactly the same every time
Yes, every computer (for several kernel versions now)
Yes, it works in Windows straight away (without doing any tricks)

One thing I have not tried yet is to set delay_use option of
usb-storage to a higher value. Maybe the device is still starting up
and kernel tries to initialize it too quickly. I have usb-storage
compiled into kernel now, instead of as a module, so I need to
remember to change that that next time I plan to reboot.

[gentoo-user] Re: USB stick recognition problem

[walt]

On 02/02/2011 03:05 PM, Paul Hartman wrote:
> On Wed, Feb 2, 2011 at 3:29 PM, walt<w41ter@gmail.com>  wrote:
>> On 02/02/2011 07:48 AM, Paul Hartman wrote:
>>
>>> I have a USB SD-card reader which cannot read the partition table when
>>> first inserted into the PC.

>> That sounds to me like a bug :)  do you see the same  on other computers?

> Yes, every computer (for several kernel versions now)

Well, if an older kernel uses the card reader normally and newer kernels
don't, then I assume a kernel bug is responsible.

What *I* would do is to use git-bisect in Linus's kernel git repository to
isolate the "bad" commit and then report it to the person who submitted the
original bad patch to Linus.

If that idea sounds weird -- I plead nolo contendere.  Yet, it gets kernel
bugs fixed.  (Very roughly paraphrasing Galileo ;)

Re: [gentoo-user] Re: USB stick recognition problem

[Paul Hartman]

On Wed, Feb 2, 2011 at 8:11 PM, walt <w41ter@gmail.com> wrote:
> On 02/02/2011 03:05 PM, Paul Hartman wrote:
>>
>> On Wed, Feb 2, 2011 at 3:29 PM, walt<w41ter@gmail.com>  wrote:
>>>
>>> On 02/02/2011 07:48 AM, Paul Hartman wrote:
>>>
>>>> I have a USB SD-card reader which cannot read the partition table when
>>>> first inserted into the PC.
>
>>> That sounds to me like a bug :)  do you see the same  on other computers?
>
>> Yes, every computer (for several kernel versions now)
>
> Well, if an older kernel uses the card reader normally and newer kernels
> don't, then I assume a kernel bug is responsible.
>
> What *I* would do is to use git-bisect in Linus's kernel git repository to
> isolate the "bad" commit and then report it to the person who submitted the
> original bad patch to Linus.
>
> If that idea sounds weird -- I plead nolo contendere.  Yet, it gets kernel
> bugs fixed.  (Very roughly paraphrasing Galileo ;)

I meant that I've tried it for a few kernel versions (it's not a new
card reader, it's a few years old). It has never worked properly in
Linux since I've owned it.

Re: [gentoo-user] Re: USB stick recognition problem

[Alan McKinnon]

Apparently, though unproven, at 08:08 on Thursday 03 February 2011, Paul 
Hartman did opine thusly:

> On Wed, Feb 2, 2011 at 8:11 PM, walt <w41ter@gmail.com> wrote:
> > On 02/02/2011 03:05 PM, Paul Hartman wrote:
> >> On Wed, Feb 2, 2011 at 3:29 PM, walt<w41ter@gmail.com>  wrote:
> >>> On 02/02/2011 07:48 AM, Paul Hartman wrote:
> >>>> I have a USB SD-card reader which cannot read the partition table when
> >>>> first inserted into the PC.
> >>> 
> >>> That sounds to me like a bug :)  do you see the same  on other
> >>> computers?
> >> 
> >> Yes, every computer (for several kernel versions now)
> > 
> > Well, if an older kernel uses the card reader normally and newer kernels
> > don't, then I assume a kernel bug is responsible.
> > 
> > What *I* would do is to use git-bisect in Linus's kernel git repository
> > to isolate the "bad" commit and then report it to the person who
> > submitted the original bad patch to Linus.
> > 
> > If that idea sounds weird -- I plead nolo contendere.  Yet, it gets
> > kernel bugs fixed.  (Very roughly paraphrasing Galileo ;)
> 
> I meant that I've tried it for a few kernel versions (it's not a new
> card reader, it's a few years old). It has never worked properly in
> Linux since I've owned it.

Then the reader itself is probably horribly broken. Or has been built to 
comply to "whatever broken Windows is doing today"

My USD card reader JustWorks(tm) everywhere with everything. And they are dirt 
cheap, about the price of the smallest SD card I can buy.

Time for a new reader perhaps?


-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] Re: USB stick recognition problem

[Dale]

Alan McKinnon wrote:
> Then the reader itself is probably horribly broken. Or has been built to
> comply to "whatever broken Windows is doing today"
>
> My USD card reader JustWorks(tm) everywhere with everything. And they are dirt
> cheap, about the price of the smallest SD card I can buy.
>
> Time for a new reader perhaps?
>
>    

I have a card reader and it works with a lot of cards.  So far it has 
worked fine.  I did have a pin to bend once but most likely my fault.  
It has Targus wrote on it.  I assume that is the brand.  I think it was 
about $10 or $15 or so.  It works fine with Linux so far and I think I 
used it once on my brothers windoze rig.

It does sound like there is something fishy about your card tho.

Dale

:-)  :-)

Re: [gentoo-user] Re: USB stick recognition problem

[Paul Hartman]

On Thu, Feb 3, 2011 at 12:52 AM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> Apparently, though unproven, at 08:08 on Thursday 03 February 2011, Paul
> Hartman did opine thusly:
>
>> On Wed, Feb 2, 2011 at 8:11 PM, walt <w41ter@gmail.com> wrote:
>> > On 02/02/2011 03:05 PM, Paul Hartman wrote:
>> >> On Wed, Feb 2, 2011 at 3:29 PM, walt<w41ter@gmail.com>  wrote:
>> >>> On 02/02/2011 07:48 AM, Paul Hartman wrote:
>> >>>> I have a USB SD-card reader which cannot read the partition table when
>> >>>> first inserted into the PC.
>> >>>
>> >>> That sounds to me like a bug :)  do you see the same  on other
>> >>> computers?
>> >>
>> >> Yes, every computer (for several kernel versions now)
>> >
>> > Well, if an older kernel uses the card reader normally and newer kernels
>> > don't, then I assume a kernel bug is responsible.
>> >
>> > What *I* would do is to use git-bisect in Linus's kernel git repository
>> > to isolate the "bad" commit and then report it to the person who
>> > submitted the original bad patch to Linus.
>> >
>> > If that idea sounds weird -- I plead nolo contendere.  Yet, it gets
>> > kernel bugs fixed.  (Very roughly paraphrasing Galileo ;)
>>
>> I meant that I've tried it for a few kernel versions (it's not a new
>> card reader, it's a few years old). It has never worked properly in
>> Linux since I've owned it.
>
> Then the reader itself is probably horribly broken. Or has been built to
> comply to "whatever broken Windows is doing today"
>
> My USD card reader JustWorks(tm) everywhere with everything. And they are dirt
> cheap, about the price of the smallest SD card I can buy.
>
> Time for a new reader perhaps?

Of course, I have others. I was just reporting to Helmut what
workaround I've used in case his has a similar problem.

The misbehaving one is extremely fast when it works, compared to the
others I've tried. It uses the Silicon Motion SM331 chipset. It's a
generic no-name reader. I've seen reports that newer packages of the
same reader now use the SM334 chipset instead, which I imagine fixes
whatever problems may have been present in the older revision.

I'm generally more interested in learning why things are broken than
using things that are not. :)

I didn't mean to derail the original thread. Sorry about that!

Re: [gentoo-user] USB stick recognition problem

[Helmut Jarausch]

On 02/02/2011 04:48:40 PM, Paul Hartman wrote:
> On Mon, Jan 31, 2011 at 3:43 AM, Helmut Jarausch
> <jarausch@igpm.rwth-aachen.de> wrote:
> > Hi,
> >
> > since a few weeks I have a strange effect with my USB stick.
> >
> > According to fdisk there is one partition on it
> > /dev/sde1              38     7839719     3919841    b  W95 FAT32
> >
> > which I haven't changed for a long time.
> >
> > Whenever I insert this stick, the kernel log shows
> > /dev/sde  but not /dev/sde1  (and there is no file /dev/sde1)
> >
> > After Invoking fdisk /dev/sde with a simple 'p' command but nothing
> > else, this device shows up.
> >
> > Has anybody an idea what's going on here?
> 
> I have a USB SD-card reader which cannot read the partition table 
> when
> first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
> the kernel re-reads the partition table and everything works fine
> after that. You can try it, maybe it'll work for you, too.

Thanks Paul and all other who tried to help.
Sorry for my late reply, I have been too busy.

My situation seems special.

- the card reader is 2 years old (builtin)
- the stick is about 1 year old

2 of my 3 USB stick work just fine

up to kernel 2.6.36-? the stick in question
has worked definitely.
Probably since kernel 2.6.37 this one shows the mentioned problems.

Paul's advice above works just fine!
So, I think it's a timing problem as Paul has suggested in a later 
mail. Where can I set the delay_use option of usb-storage to a higher 
value?

Many thanks again,
Helmut.

Re: [gentoo-user] USB stick recognition problem

[Paul Hartman]

On Fri, Feb 4, 2011 at 8:48 AM, Helmut Jarausch
<jarausch@igpm.rwth-aachen.de> wrote:
> On 02/02/2011 04:48:40 PM, Paul Hartman wrote:
>> On Mon, Jan 31, 2011 at 3:43 AM, Helmut Jarausch
>> <jarausch@igpm.rwth-aachen.de> wrote:
>> > Hi,
>> >
>> > since a few weeks I have a strange effect with my USB stick.
>> >
>> > According to fdisk there is one partition on it
>> > /dev/sde1              38     7839719     3919841    b  W95 FAT32
>> >
>> > which I haven't changed for a long time.
>> >
>> > Whenever I insert this stick, the kernel log shows
>> > /dev/sde  but not /dev/sde1  (and there is no file /dev/sde1)
>> >
>> > After Invoking fdisk /dev/sde with a simple 'p' command but nothing
>> > else, this device shows up.
>> >
>> > Has anybody an idea what's going on here?
>>
>> I have a USB SD-card reader which cannot read the partition table
>> when
>> first inserted into the PC. But, if I issue "hdparm -z /dev/sdg" then
>> the kernel re-reads the partition table and everything works fine
>> after that. You can try it, maybe it'll work for you, too.
>
> Thanks Paul and all other who tried to help.
> Sorry for my late reply, I have been too busy.
>
> My situation seems special.
>
> - the card reader is 2 years old (builtin)
> - the stick is about 1 year old
>
> 2 of my 3 USB stick work just fine
>
> up to kernel 2.6.36-? the stick in question
> has worked definitely.
> Probably since kernel 2.6.37 this one shows the mentioned problems.
>
> Paul's advice above works just fine!
> So, I think it's a timing problem as Paul has suggested in a later
> mail. Where can I set the delay_use option of usb-storage to a higher
> value?

I think you must first ensure usb-storage is built as a module, then
create a file /etc/modprobe.d/usb_storage with these contents:

options usb_storage delay_use=5

That would set the delay to 5 seconds. And you may need to run
update-modules after changing that file.

I haven't tried it yet and that's based on my memory as I'm using a
Windows computer at the moment... Good luck!

Re: [gentoo-user] USB stick recognition problem

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 622 bytes --]

On Fri, 4 Feb 2011 09:52:12 -0600, Paul Hartman wrote:

> > So, I think it's a timing problem as Paul has suggested in a later
> > mail. Where can I set the delay_use option of usb-storage to a higher
> > value?  
> 
> I think you must first ensure usb-storage is built as a module, then
> create a file /etc/modprobe.d/usb_storage with these contents:
> 
> options usb_storage delay_use=5

You can set options for built in modules too, by adding them to the
kernel line in your boot loader, e.g.

usb_storage.delay_use=5


-- 
Neil Bothwick

Celery is not food. It is a member of the plywood family.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]