Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

lilsirecho · 2011-12-19 04:22:58

Booted into xfce4 Desktop with mouse cursor in center of display and not enabled. Could not use the keybd. Had to hard reset.

Perhaps this is an xfce4 problem . X is ok AFAICT since the display has full icons and a background displayed jpg image. This is setup within xfce4 so that xfce4 doesn't seem to be the problem.

Have modified the mkinitcpio.conf for usbinput as provided by .pacnew.

Perhaps I forget something?

Maybe something in rc.conf?

Brain strain at 86 years old!!!!

Last edited by lilsirecho (2012-01-05 16:13:41)

karol · 2011-12-21 16:42:33

Did you fix it? If not, can you post some logs?

lilsirecho · 2011-12-22 03:22:37

karol:

Haven;t had time to spend on it so I haven't fixed it. I think I have a kernel problem, though and grub needs attention.

Have to wait until after Xmas to hit it again.

Doing well with whatever though altho haven't thot of any good new jokes..

Liked the one about higgs boson and mass!!

samuvuo · 2011-12-25 06:05:40

Some things to check:

- Do you have packages xf86-input-evdev, xf86-input-keyboard and xf86-input-mouse installed?
- Do you have /etc/X11/xorg.conf.d/10-evdev.conf?
- Do you have /etc/X11/xorg.conf (or similar) that could interfere?

lilsirecho · 2011-12-25 17:33:35

All the listed items are installed.

lilsirecho · 2011-12-29 16:56:22

Karol:
Appreciate your comment and unload my problem more completely in this post This problem is a raid0 related problem introduced when upgrading to linux3.1.5-1 on my bootable flash-based raid0 archlinux.

While in chroot to /dev/md0: (Booted into linux3.1.5-1 in HDD)

[root@n6re /]# lsmod
Module                  Size  Used by
ext2                   63283  1 
raid0                   7958  1 
snd_hda_codec_realtek   282272  1 
i915                  725922  2 
snd_hda_intel          22410  0 
snd_hda_codec          81242  2 snd_hda_codec_realtek,snd_hda_intel
drm_kms_helper         25721  1 i915
i2c_i801                8187  0 
pcspkr                  1819  0 
evdev                   9530  11 
drm                   185768  3 i915,drm_kms_helper
ppdev                   5774  0 
i2c_algo_bit            5199  1 i915
i2c_core               20460  5 i915,drm_kms_helper,i2c_i801,drm,i2c_algo_bit
video                  11164  1 i915
processor              24224  0 
parport_pc             30042  0 
parport                31631  2 ppdev,parport_pc
r8169                  45411  0 
iTCO_wdt               11789  0 
mii                     4059  1 r8169
iTCO_vendor_support     1961  1 iTCO_wdt
button                  4470  1 i915
intel_agp              10904  1 i915
intel_gtt              14455  3 i915,intel_agp
ext3                  191948  3 
jbd                    63179  1 ext3
mbcache                 5881  2 ext2,ext3
md_mod                100392  2 raid0
sd_mod                 28275  9 
usbhid                 35352  0 
hid                    82435  1 usbhid
pata_acpi               3376  0 
ata_piix               22077  2 
uhci_hcd               23084  0 
sata_sil24             11683  4 
libata                166724  3 pata_acpi,ata_piix,sata_sil24
scsi_mod              132826  2 sd_mod,libata
ehci_hcd               40762  0 
snd_usb_audio          93005  0 
snd_pcm                74368  3 snd_hda_intel,snd_hda_codec,snd_usb_audio
snd_timer              19544  1 snd_pcm
snd_page_alloc          7153  2 snd_hda_intel,snd_pcm
snd_hwdep               6357  2 snd_hda_codec,snd_usb_audio
snd_usbmidi_lib        18697  1 snd_usb_audio
usbcore               144432  6 usbhid,uhci_hcd,ehci_hcd,snd_usb_audio,snd_usbmidi_lib
snd_rawmidi            19458  1 snd_usbmidi_lib
snd_seq_device          5300  1 snd_rawmidi
snd                    58362  10 snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_usb_audio,snd_pcm,snd_timer,snd_hwdep,snd_usbmidi_lib,snd_rawmidi,snd_seq_device
soundcore               6210  1 snd
[root@n6re /]#

Again with chroot:

[root@n6re /]# ls dev
agpgart		    ppp       tty20  tty55    vcs25  vcs6    vcsa36
autofs		    psaux     tty21  tty56    vcs26  vcs60   vcsa37
block		    ptmx      tty22  tty57    vcs27  vcs61   vcsa38
bsg		    pts       tty23  tty58    vcs28  vcs62   vcsa39
btrfs-control	    random    tty24  tty59    vcs29  vcs63   vcsa4
bus		    rtc       tty25  tty6     vcs3   vcs7    vcsa40
char		    rtc0      tty26  tty60    vcs30  vcs8    vcsa41
console		    sda       tty27  tty61    vcs31  vcs9    vcsa42
core		    sda1      tty28  tty62    vcs32  vcsa    vcsa43
cpu		    sda2      tty29  tty63    vcs33  vcsa1   vcsa44
cpu_dma_latency     sda3      tty3   tty7     vcs34  vcsa10  vcsa45
disk		    sda4      tty30  tty8     vcs35  vcsa11  vcsa46
dri		    sdb       tty31  tty9     vcs36  vcsa12  vcsa47
fb0		    sdb1      tty32  ttyS0    vcs37  vcsa13  vcsa48
fd		    sdb2      tty33  ttyS1    vcs38  vcsa14  vcsa49
full		    sdc       tty34  ttyS2    vcs39  vcsa15  vcsa5
fuse		    sdc1      tty35  ttyS3    vcs4   vcsa16  vcsa50
hidraw0		    sdc2      tty36  uinput   vcs40  vcsa17  vcsa51
hidraw1		    shm       tty37  urandom  vcs41  vcsa18  vcsa52
hidraw2		    snapshot  tty38  usb      vcs42  vcsa19  vcsa53
hpet		    snd       tty39  vcs      vcs43  vcsa2   vcsa54
initctl		    stderr    tty4   vcs1     vcs44  vcsa20  vcsa55
input		    stdin     tty40  vcs10    vcs45  vcsa21  vcsa56
kmsg		    stdout    tty41  vcs11    vcs46  vcsa22  vcsa57
log		    tty       tty42  vcs12    vcs47  vcsa23  vcsa58
loop-control	    tty0      tty43  vcs13    vcs48  vcsa24  vcsa59
loop0		    tty1      tty44  vcs14    vcs49  vcsa25  vcsa6
mapper		    tty10     tty45  vcs15    vcs5   vcsa26  vcsa60
mcelog		    tty11     tty46  vcs16    vcs50  vcsa27  vcsa61
md		    tty12     tty47  vcs17    vcs51  vcsa28  vcsa62
md0		    tty13     tty48  vcs18    vcs52  vcsa29  vcsa63
mem		    tty14     tty49  vcs19    vcs53  vcsa3   vcsa7
net		    tty15     tty5   vcs2     vcs54  vcsa30  vcsa8
network_latency     tty16     tty50  vcs20    vcs55  vcsa31  vcsa9
network_throughput  tty17     tty51  vcs21    vcs56  vcsa32  vga_arbiter
null		    tty18     tty52  vcs22    vcs57  vcsa33  watchdog
parport0	    tty19     tty53  vcs23    vcs58  vcsa34  zero
port		    tty2      tty54  vcs24    vcs59  vcsa35
[root@n6re /]#

This indicates that /dev/md0 is extant in my hdd boot-up in Linux3.1.5-1 x86_64.

The problem I face is as follows:

I have generated a unique bootable raid0 system utilizing a setup derived from 2010-05 .iso wherein I generated /dev/md0 as root and a partitioned boot. This was generated from FTP install so it had kernel26v39.

This performed very nicely using jumanji or browsing and booted in nine seconds to xfce4.

Recently I made a copy of the /dev/md0 raid array and upgraded it to Linux3..5-1 x96_64. Therein lies the problem.

When I run the normal mdadm setup for mdadm.conf which is done with the following:

mdadm -D --scan >>/etc/mdadm.conf...

The result is the /dev/md0 is no longer referenced in mdadm.conf but the component boot device uuid is listed.

Proc shows the following:

[root@n6re /]# cat /proc/mdstat
Personalities : [raid0] 
md0 : active raid0 sdc2[1] sdb2[0]
      31016960 blocks super 1.2 512k chunks

Thus when booting into the new upgrade, the boot device is recognized as /dev/md0 by uuid and the boot sequence is performed and reaches xfce4 Desktop.

At this point, neither keyboard nor mouse is active and a hard reset is needed.

When utilizing gparted to analyze the devices, some indication of what is happening occurs. It reports that /dev/md/0..(.that's correct /md/0 ) is not a valid device.

It therefore seems that Linux3 kernel changes the ID of /dev/md0 to /dev/md/0 by some activity such as udev,mdadm,or raid conventions.

The grub version is use for all the raid0 devices is version 1.99.

Please appreciate that I have been advised to not run root only and not run flash devices in raid0. My experience with running as root is now 9 years. As to raid0 use with flash devices....within the device algorithms are provided for wearleveling and bad block controls (algorithms probably different for each manufacturer) but the key to using flash devices in raid0 is to provide many more cells than are needed for activities such that the same cells are never written over and over and to provide more ram than can reasonably be needed for activities to prevent swapping.

I appreciate the speed of boot and the jumanji browser speed and xfce4 simplicity because I am 86 years old and don't have time for slow loading and slow access...so there!

I am attempting to solve this problem today by re-nstalling mkinitcpio.conf and running mkinitcpio to see if that provides keybd and mouse.

I am flabbergasted that the boot to Desktop occurs without the root device /md0 being listed in mdadm.conf!!!! Too many irons in the fire since the system proceeds without mdadm.conf being arrayed at all!!!!!

This indicates that a user has poor control over a raid system.

Essentially all of the googling for info applies to raid1,5,6 or 10. Practically nothing on raid0, especially booting from raid0.

I have successfully used --grow on my raid0 to increase the size of my raid pair.

I have not been able to use --grow with --add to raise the number of devices as is allowed in mdadm man pages. The application requires a move to raid4 to resync and then return to raid0. It refuses to move to level4. This attempt was made while booted in HDD to linux3.1.5-1. Thus it may be related to the handling of this function in that kernel.

Will try the mkinitcpio activity and report back.

lilsirecho · 2011-12-30 21:23:04

Installed mkinitcpio.conf.pacnew but the problem persists.

Have reached Desktop in xfce4 with everything displayed but cannot use ketbd nor mouse.

Suspect it may be due to xorg vmmouse in X11.

lilsirecho · 2011-12-31 16:55:22

Remodeled grub to accept the new kernel and lost the boot capability into desktop. It was using the wrong grub elements.

Now have lost the ability to boot the root device /dev/md0 as is usual with the new linux kernel. It seems not to recognize the root device as /dev/md0 .

Essentially the posting is now moot.

lilsirecho · 2012-01-04 15:20:51

After much study of the problem, I have developed a "fix" for the boot in Linux 3. EDIT: Now booting in linux-3.1.6-1-x86_64

At the grub prompt change the set root= to (hd1,1).....

Add root=/dev/md0 to the linux kernel line....

Make the initramfs change.

The system then boots into xfce4 Desktop in nine seconds.

Problem partially solved....need to fix grub.cfg...................

How?

Last edited by lilsirecho (2012-01-04 15:35:22)

lilsirecho · 2012-01-04 18:35:31

Tried editing grub.cfg with the changes which allowed booting into system.

This did not allow booting...failed to load kernel, even though the changes appeared in the edit grub window.
What other procedure is necessary to enable booting?

lilsirecho · 2012-01-04 18:54:35

Further info:

There is no grub file in /boot directory.

The grub file appears in the first partition of raid0 devices and is selected in bios for booting.
EDIT: Restoring grub to original form and editing the prompt again allows for system to reach Desktop.

Last edited by lilsirecho (2012-01-04 19:13:08)

lilsirecho · 2012-01-04 19:21:04

A second raid0 pair of 533x adata 16gb devices produces the same result as the maxell devices except for the hdparm results for /dev/md0....

sh-4.2# hdparm -t /dev/md0

/dev/md0:
 Timing buffered disk reads: 378 MB in  3.01 seconds = 125.51 MB/sec
sh-4.2# hdparm -t /dev/md0

/dev/md0:
 Timing buffered disk reads: 378 MB in  3.01 seconds = 125.46 MB/sec
sh-4.2# 
sh-4.2# hdparm -t /dev/md0

/dev/md0:
 Timing buffered disk reads: 378 MB in  3.01 seconds = 125.63 MB/sec
sh-4.2#

The results for the maxell devices is ~180MB/s.

This shows differences in internal algorithms (perhaps) in manufacturing.

lilsirecho · 2012-01-05 02:37:20

Following the suggestion that set root= (hd1,1) means what it says, I understand why it now boots in raid0 with that set root and /devmd0 added to the kernel line.

It seems that the HDD is defined as hd1 since it is the second drive listed in bios. Thus it starts the sequence until mdadm inserts....." /dev/md0 with 2 drives" and from then on root is defined by the added /dev/mdo in the kernel line.

This is certain since the display at Desktop is correct only for the raid array and is not that of the HDD.

If I change the position of the HDD in bios so it is not the second drive therein, the system does not reach kernel stage in boot and fails to boot up in raid0.

So HDD acts as a helper in setting up /dev/md0 which does not occur without it in the linux 3 kernels.

The real culprit is the combo of udev, mkinitcpio, mdadm and the kernel which do not permit /dev /md0 to be recognized in a raid0 partition boot.

Device HDD is a four partition device, boot, swap,root, and home.

Raid0 is only boot, swap and root, boot being 100mb and root 29.6mb. Swap is also 100mb.

Thus, the raid0 continues to be elusive in linux 3 kernels when using partitioned grub and partitioned root with /dev/md0.

I am pleased that I can get it going but still chagrined at how it must be done!

lilsirecho · 2012-01-05 02:56:33

Another strange thing: the array devices aren't identified correctly in /etc/mdadm.conf. The uuid of md0 is assigned as that of the component drives rather than the correct /dev/md0 assignment.

ARRAY /dev/md0 metadata=1.2 name=(none):0 UUID=476e7e0a:fb08db59:d2592e55:73d19$
ARRAY /dev/md0 metadata=1.2 name=(none):0 UUID=476e7e0a:fb08db59:d2592e55:73d19$
ARRAY /dev/md1 metadata=1.2 name=(none):0 UUID=476e7e0a:fb08db59:d2592e55:73d19$

Thus it would seem that the use of mdadm.conf might cause the failed raid0 boot.

I will attempt to change the array to reflect the uuid for /dev/md0 which is utilized in grub boot partition. It may permit the system to boot........

EDIT: Change did not solve the raid0 /dev/md0 boot problem. It still requires the helper HDD.

Last edited by lilsirecho (2012-01-05 03:24:51)

lilsirecho · 2012-01-05 04:15:02

It seems obvious that the kernel does not utilize mdadm.conf to generate raid arrays since that is the only thing missing from the boot sequence as displayed....until the HDD does the mdadm recognition of mdadm.conf data.

Watching the progress of the boot-up via the led indicators shows the CF cards responding to the booting sequence following the HDD identification of /dev/md0.

This is supported by the Desktop display of raid0 pertinent data when booting is complete. /dev/md0 is listed in ls dev and in blkid. Therefore the device exists ( and in fact is mounted in /mnt/md).

Thus it seems imperative that the linux kernel recognize mdadm.conf and initiate root containing the OS.

Perhaps this can be arranged? What params can be entered to the kernel line to effect this required activity? The first activity seems to be udev and it might be the culprit in defining the handling of mdadm.conf by misassigning the uuiid for /dev/md0.

Kinda messy this!

EDIT:Upgrade is now 3.1.6-1 with this problem.

Last edited by lilsirecho (2012-01-05 04:19:05)

lilsirecho · 2012-01-05 16:11:49

Upgraded to linux 3.1.7-1 and the boot reached Desktop with no keybd or mouse control.

Therefore, upgraded the HDD to linux 3.1.7-1 and rebooted in raid0 system

Reached Desktop and keybd and mouse active.

As long as I have the HDD for backup, the raid0 system can function.EDIT: The HDD has to have the same kernel version as the upgrade to the raid0 array.!!!!!!! This also works in like manner from USB HDD of same kernel version after which the USB HDD can be powered off to save power.

Not unlike using a floppy for grub booting.

Closing this post.

Last edited by lilsirecho (2012-01-07 03:54:33)

Arch Linux

#1 2011-12-19 04:22:58

Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#2 2011-12-21 16:42:33

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#3 2011-12-22 03:22:37

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#4 2011-12-25 06:05:40

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#5 2011-12-25 17:33:35

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#6 2011-12-29 16:56:22

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#7 2011-12-30 21:23:04

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#8 2011-12-31 16:55:22

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#9 2012-01-04 15:20:51

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#10 2012-01-04 18:35:31

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#11 2012-01-04 18:54:35

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#12 2012-01-04 19:21:04

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#13 2012-01-05 02:37:20

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#14 2012-01-05 02:56:33

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#15 2012-01-05 04:15:02

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

#16 2012-01-05 16:11:49

Re: Upgrade to 3.1.7-1 : no usb keybd or mouse(Workaround solution)

Board footer