hotplug on different hardware

[johnbock]

Being that I have only the 7270, I'm wondering how the directory /sys differ to mine.
Hoping someone could help me out here: can you please do the following

ls -lR /sys >/tmp/$(uname -r)_sysfs.list
find /sys -name dev >/tmp/$(uname -r)_sysfs.find

Please write the box model in your post and attach the compressed files /tmp/*_sysfs.*.

Thank you all in advance:groesste:


Now we have:
7270
7170
7141
W900V

Are the others like the 7170, 7141, W900V?

 

[poruid]

Here the requested listings of a 7170 that has one memory stick inserted that is used for usbroot.

 

[wengi]

Hi,

7170 and 7141 both with USB Stick attached (Data only, no USBroot)

wengi

EDIT: i also have a W701V with 7170 alien firmware if you need this too.

 

[johnbock]

Thanks for the information!

Oh, that looks bad, the ahci-hcd support for /sys is only half done: no '/dev', no endpoint.

I may have an idea... Could you search for modalias

find /sys -name modalias

and attach please.

 

[matze1985]

on a 7141

# find /sys -name modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1.2/1-1.2:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1.4/1-1.4:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1.3/1-1.3:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-0:1.0/modalias

 

[Silent-Tears]

I think I should give you the data for my W900V, too. It build with alien hardware and uses a nearly actual trunk.

 

[Darkyputz]

on a 7170

/sys/devices/ahci-hcd/usb1/1-1/1-1.1/1-1.1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-0:1.0/modalias

 

[poruid]

I may have an idea... Could you search for modalias

find /sys -name modalias

and attach please.

With cat output of the files found, whatever purpose it may serve.

/var/mod/root # find /sys -name modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-0:1.0/modalias
/var/mod/root # find /sys -name modalias -exec cat {} \;
usb:v0781p5406d0200dc00dsc00dp00ic08isc06ip50
usb:v0000p0000d0000dc09dsc00dp00ic*isc*ip*

 

[wengi]

7141:

/var/mod/root # find /sys -name modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-0:1.0/modalias
/var/mod/root #

7170:

/var/mod/root # find /sys -name modalias
/sys/devices/ahci-hcd/usb1/1-1/1-1:1.0/modalias
/sys/devices/ahci-hcd/usb1/1-0:1.0/modalias

again both with USB Stick (data only) attached.
wengi

 

[MrXYZ]

7270 16MB with usb hub and CF Cardreader/256MB CF Card

/var/mod/root # find /sys -name modalias
/sys/devices/pci0000:00/0000:00:0e.0/modalias
/sys/devices/platform/musb_hdrc/usb1/1-1/1-1.2/1-1.2:1.0/modalias
/sys/devices/platform/musb_hdrc/usb1/1-1/1-1:1.0/modalias
/sys/devices/platform/musb_hdrc/usb1/1-0:1.0/modalias
/sys/devices/platform/musb_hdrc/modalias
/sys/devices/platform/serial8250/modalias
/sys/devices/platform/mtd-ram.2/modalias
/sys/devices/platform/physmap-flash.1/modalias

 

[johnbock]

Hi,

Can someone test the two attached patches in combination with hotplug II / mdev implementation on a 7141, W900V or 7170. These attachments must be unpacked to make/busybox/patches and make busybox-dirclean.

Although I have tested this on the 7270, I'm not sure how the 2.6.13 based boxes will react. Meaning, only technically adverse people should test this. I believe, the worst that can happen is that a make recover must be performed, so don't forget to backup your configuration before testing. DONOT test this with any boot options, such as usb_root, mini_fo or the like.

Once the box is up and running, login and do:

mdev -s
ls -lRtr /dev

and please post the results. Oh and please don't forget to tell me on what model it was tested.

 

[johnbock]

Has anyone tested these with hp II?

Do you need more infomation?

 

[wengi]

Hi,

i am sorry, but the boxes for testing are W701V. But there is no USB

wengi

 

[poruid]

Has anyone tested these with hp II?

Had the VM in which the firmware is build not crashed beyond repair, your patches would have been tested already; sorry, but recreating the VM has so far proved difficult. Hopefully the test can be performed this night.

OT: after some progressively worsening crashes of the VMWare Fusion 2 beta hosted VM and unsuccesfull recreations of the ubuntu VM (crashes during install, corrupted FS beyond repair with fsck), I've decided to remove the beta and reinstall the 1.1.3 releases, in order to recreate the ubuntu VM.

 

[johnbock]

Hopefully the test can be performed this night.

That's unfortunate... But thanks for taking the time and informing me of your intent. Take your time, I can wait

 

[poruid]

testresult for 7170 AVM 29.04.57, freetz 2375

mdev seems to work without problems, see extract from the syslog.
Fast remarks:
- when umount is run from the command line, then the devmap file is not updated.
- after replugging the memory key, the FS is setup as /dev/sdb...,
/dev/sdb1 is mounted on /var/media/ etc.
/dev/sdb2 is added to the swap space
- the "Geräteübersicht" shows the mounts, the swap partition as well!
USB-Gerät Aktivierte Geräteoptionen
uStorb2/swap USB-Netzwerkspeicher
/var/media/ftp/uStorb1 USB-Netzwerkspeicher
- cold plugging works.

If this hotplug implementation would also work with usb-root ...
Anyway, chapeau.

Jul 15 22:57:24 fritz user.notice kernel:   Vendor: SanDisk   Model: U3 Cruzer Micro   Rev: 6.51
Jul 15 22:57:24 fritz user.notice kernel:   Type:   Direct-Access                      ANSI SCSI revision: 00
Jul 15 22:57:24 fritz user.notice kernel: SCSI device sda: 2014271 512-byte hdwr sectors (1031 MB)
Jul 15 22:57:24 fritz user.notice kernel: sda: Write Protect is off
Jul 15 22:57:24 fritz user.debug kernel: sda: Mode Sense: 45 00 00 08
Jul 15 22:57:24 fritz user.err kernel: sda: assuming drive cache: write through
Jul 15 22:57:24 fritz user.notice kernel: SCSI device sda: 2014271 512-byte hdwr sectors (1031 MB)
Jul 15 22:57:24 fritz user.notice kernel: sda: Write Protect is off
Jul 15 22:57:24 fritz user.debug kernel: sda: Mode Sense: 45 00 00 08
Jul 15 22:57:24 fritz user.err kernel: sda: assuming drive cache: write through
Jul 15 22:57:24 fritz user.info kernel:  sda: sda1 sda2
Jul 15 22:57:24 fritz user.notice kernel: Attached scsi removable disk sda at scsi0, channel 0, id 0, lun 0
Jul 15 22:57:24 fritz user.notice kernel: Attached scsi generic sg0 at scsi0, channel 0, id 0, lun 0,  type 0
Jul 15 22:57:24 fritz user.notice kernel:   Vendor: SanDisk   Model: U3 Cruzer Micro   Rev: 6.51
Jul 15 22:57:24 fritz user.notice kernel:   Type:   CD-ROM                             ANSI SCSI revision: 00
Jul 15 22:57:24 fritz user.notice kernel: Attached scsi generic sg1 at scsi0, channel 0, id 0, lun 1,  type 5
Jul 15 22:57:24 fritz user.debug kernel: usb-storage: device scan complete
Jul 15 22:57:24 fritz user.warn kernel: AR7WDT: System Init UEberwachung abgeschlossen (196610 ms noch verfuegbar)
Jul 15 22:57:24 fritz user.info kernel: Adding 128512k swap on /dev/sda2.  Priority:-1 extents:1
...
Jul 15 22:57:49 fritz local0.info partition[822]: /dev/sda1 was not cleanly unmounted, check forced. /dev/sda1: 4350/54880 files (0.1% non-contiguous), 27580/218877 blocks
Jul 15 22:57:49 fritz local0.info partition[822]: mounted '/dev/sda1' on '/var/media/ftp/uStora1' with 'noatime,nodiratime'
Jul 15 22:57:49 fritz local0.info partition[822]: not starting any services on '/var/media/ftp/uStora1'
...

 

[johnbock]

mdev seems to work without problems, see extract from the syslog.
When this hotplug implementation would also work with usb-root ...
Anyway, chapeau.

...
Jul 15 22:57:49 fritz local0.info partition[822]: /dev/sda1 was not cleanly unmounted, check forced. /dev/sda1: 4350/54880 files (0.1% non-contiguous), 27580/218877 blocks
Jul 15 22:57:49 fritz local0.info partition[822]: mounted '/dev/sda1' on '/var/media/ftp/uStora1' with 'noatime,nodiratime'
Jul 15 22:57:49 fritz local0.info partition[822]: not starting any services on '/var/media/ftp/uStora1'
...

Only this portion of your log is from mdev... And there still seems to be something missing...

Using the avm web interface disable the memory key and disconnect it from the box. And could you kindly post the results of

ls -altrR /dev/
mdev -s
ls -altrR /dev/
lsmod

and the appropriate portion of your log please?

TIA

 

[johnbock]

- when umount is run from the command line, then the devmap file is not updated.

/lib/mdev/interface/storage unplug /var/media/ftp<entire mountpath>

The prefixing /var/media/ftp is a work-around for the avm web interface. In the next version (0.0.5) the prefix will not be necessary when called from the command line

If this hotplug implementation would also work with usb-root ...

It shall, all in due time... Could you kindly reply to the above post?

 

[poruid]

Only this portion of your log is from mdev... And there still seems to be something missing...

Using the avm web interface disable the memory key and disconnect it from the box. And could you kindly post the results of

ls -altrR /dev/
mdev -s
ls -altrR /dev/
lsmod

and the appropriate portion of your log please?
TIA

The tests I've run were:

<booted without memory key inserted>
ls -altrR /dev/ >before-insert.lsdev
lsmod >before-insert.lsmod

<memory inserted>
ls -altrR /dev/ >after-insert.lsdev
lsmod >after-insert.lsmod

<entfernen in AVM GUI>
ls -altrR /dev/ >after-entfernen.lsdev
lsmod >after-entfernen.lsmod

<disconnect memory key>
ls -altrR /dev/ >after-disconnect.lsdev
lsmod >after-disconnect.lsmod

<run mdev -s>
ls -altrR /dev/ >after-mdev-s.lsdev
lsmod >after-mdev-s.lsmod

<re-insert memory key>
ls -altrR /dev/ >after-reinsert.lsdev
lsmod >after-reinsert.lsmod

I suppose that this yields a pretty complete picture of what happens, see the attached tar file with the test outputs and relevant extracts from the syslog.
Good luck.

 

[poruid]

Zitat von poruid Beitrag anzeigen
- when umount is run from the command line, then the devmap file is not updated.

This behaviour has not been seen since; maybe it happened once due to some unlnown circumstances or state of the FB.