dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 2:31 pm

dom wrote:Cool. I'm expecting to bump normal rpi-update tree to 3.10 this weekend.
This is now done.

Code: Select all

sudo rpi-update
will now get you the 3.10.18 kernel. Let me know if it works okay for you.

mcgyver83
Posts: 358
Joined: Fri Oct 05, 2012 11:49 am

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 3:37 pm

Great, I'll try it tonight.
Please can you provide me the changelog link for this last fw version?
About CMA (dynamic memory splitting)?

ne3e
Posts: 10
Joined: Fri Apr 20, 2012 2:09 pm

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 6:34 pm

I just ran the update on my PI Rev B 256Mb.
It updated the Kernel with no issues. Thank you developers for your fine work.

John

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 6:46 pm

mcgyver83 wrote:Great, I'll try it tonight.
Please can you provide me the changelog link for this last fw version?
About CMA (dynamic memory splitting)?
The change for the last update is just to the kernel, and it is just upstream changes. See:
https://git.kernel.org/cgit/linux/kerne ... s/v3.10.18

Pi specific patches are here (but they have been in the previous kernel)
https://github.com/raspberrypi/linux/commits/rpi-3.10.y

firmware changelog is here (but again, the gpu code is the same as it was last week)
https://github.com/raspberrypi/firmware/commits/master

No change to CMA. It's considered experimental and is not actively supported.

mcgyver83
Posts: 358
Joined: Fri Oct 05, 2012 11:49 am

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 7:32 pm

Many thanks!

B.Jay
Posts: 28
Joined: Thu Jul 18, 2013 10:57 pm

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 9:38 pm

@dom: I got a problem here with 3.10 I can't seem to resolve ...

I finally got around to make a modification (previously everything ran off a 32GB SD which I need for something else):

The RasPi boots the firmware off a 2GB SD - only contains a single FAT32 partition with the firmware files (which is then mounted under /boot) - to fire up the rootfs ("/") from sda1 (USB connected 80GB hard drive, 1 ext4 partition).

Now, upon powering on the kernel loads up but stalls at the point where it finds the USB drive (seems "mount" is sitting there waiting forever for "/dev/sda1" to become available).

I hooked up the serial debug cable to capture the output to copy/paste it here ... the last few lines are:
[ 4.672720] scsi 0:0:0:0: Direct-Access Seagate ST380815AS PQ: 0 ANSI: 4
[ 4.686998] sd 0:0:0:0: [sda] 156312576 512-byte logical blocks: (80.0 GB/74.5 GiB)
[ 4.702478] sd 0:0:0:0: [sda] Write Protect is off
[ 4.712059] sd 0:0:0:0: [sda] Mode Sense: 23 00 00 00
[ 4.712634] sd 0:0:0:0: [sda] No Caching mode page present
[ 4.722829] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 4.736362] sd 0:0:0:0: [sda] No Caching mode page present
[ 4.746588] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 6.225165] sda: sda1
[ 6.235174] sd 0:0:0:0: [sda] No Caching mode page present
[ 6.245409] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 6.256050] sd 0:0:0:0: [sda] Attached SCSI disk

The next line where mount would say that it got mounted never happens.

cmdline.txt on the SD reads:
dwc_otg.lpm_enable=0 console=ttyAMA0,115200 kgdboc=ttyAMA0,115200 console=tty1 root=/dev/sda1 rootfstype=ext4 elevator=deadline rootwait

I took another SD card and put the firmware files from 3.6.11+ on it, copied the 3.6.11+ modules and stuff to "rootfs" (sda1) ... and it boots up like a charm.

Do I miss something obvious that changed with 3.10?

EDIT: Just to clarify "why does he need to run the system off a hard drive"...

I'm running Apache and MySQL (for TT-RSS), so using a hard drive is the way smarter option than running it off the SD card (LOTS and LOTS of Read/Writes - in terms of MySQL).

Also, it's not the fault of /etc/fstab ... the file is fine, otherwise 3.6.11+ wouldn't boot up with no problems.

Another EDIT: HERE at viewtopic.php?f=66&t=60158 you have another user with the exact same problem (rootfs on a hard drive, boot fails horribly with the 3.10 update). Am I glad I don't see ghosts nor am the only one having the problem.
Last edited by B.Jay on Tue Nov 05, 2013 10:09 pm, edited 1 time in total.

JerryPi
Posts: 17
Joined: Tue Nov 20, 2012 8:25 am

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 10:07 pm

I would like to see this included (is in master and 3.9) as well:
CONFIG_R8712U=m
(Supports USB devices based on the RTL8188SU, RTL8191SU and RTL8192SU chips).
Thank you!
Don't grow up, it's a trap!

RobHenry
Posts: 452
Joined: Fri Sep 21, 2012 9:04 pm
Location: UK

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 11:35 pm

This failed to complete boot for me, at least until the point where an ethernet connection established. I'm running headless with the rootfs on an external HDD so don't know exactly what was going on; the lights looked active but no network connection was established.

The relevant section from the syslog is here: http://paste.debian.net/64203/ - it just went quiet after 22:28:26, until 23:09:54 when I got it to boot on previous firmware/kernel.

cmdline.txt:

Code: Select all

dwc_otg.lpm_enable=0 console=ttyAMA0,115200 kgdboc=ttyAMA0,115200 console=tty1 root=/dev/sda5 rootfstype=ext4 elevator=deadline rootwait
config.txt (excluding comments):

Code: Select all

arm_freq=1050
core_freq=450
sdram_freq=500
over_voltage=6
gpu_mem=16
force_turbo=1
hdmi_ignore_cec_init=1
I only waited 3 - 4 minutes before deciding it wasn't going to boot. System was recovered by plugging the sd card and USB HDD into a laptop and using rpi-update to revert to the commit from 18th October (8234d5148aded657760e9ecd622f324d140ae891).

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Tue Nov 05, 2013 11:40 pm

RobHenry wrote:This failed to complete boot for me, at least until the point where an ethernet connection established. I'm running headless with the rootfs on an external HDD so don't know exactly what was going on; the lights looked active but no network connection was established.
Sounds like this:
http://www.raspberrypi.org/phpBB3/viewt ... 66&t=60158

User avatar
g7ruh
Posts: 67
Joined: Mon Apr 23, 2012 9:49 am
Location: Blackfield UK

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 1:53 pm

Dom,
I upgraded yesterday and found that my SPI link to an A to D converter stopped working. I puzzled over it until I looked at the last update to xively.com (search G7RUH if interested) from devpi (my development system) The live one will not be updated until 'operational proving' tests are OK on the dev pi! This is the error I get when running the python job from the latest update:

Code: Select all

Traceback (most recent call last):
  File "mod3atod.py", line 33, in <module>
    spi.open(0,0)          # This is on SPI channel 0 (CE0 - aka GPIO8)
IOError: [Errno 2] No such file or directory
On reverting to the previous firmware update:

Code: Select all

sudo rpi-update 8234d5148aded657760e9ecd622f324d140ae891
everything is normal and the python job runs fine.

I then did the rpi-update again and the fault returned. I therefore conclude that something in the latest code has "broken" it, or, at least, changed the way it works.

Having just reverted to the previous version (again) all works normally.

Something strange going on! Any ideas or suggestions?

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 1:57 pm

g7ruh wrote:Dom,
Something strange going on! Any ideas or suggestions?
Probabably a kernel module is missing. Can you run your SPI program and report output of lsmod, in the working and non-working cases?

User avatar
g7ruh
Posts: 67
Joined: Mon Apr 23, 2012 9:49 am
Location: Blackfield UK

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 2:09 pm

Dom,

that was quick!

lsmod (previous firmware) python job runs OK

Code: Select all

Module                  Size  Used by
rtc_ds1307              7042  0
fuse                   69438  3
w1_therm                2995  0
w1_gpio                 1633  0
wire                   24973  2 w1_gpio,w1_therm
cn                      4834  1 wire
8192cu                490361  0
i2c_dev                 5594  18
snd_bcm2835            16432  0
snd_pcm                77728  1 snd_bcm2835
snd_page_alloc          5169  1 snd_pcm
snd_seq                53482  0
snd_seq_device          6462  1 snd_seq
snd_timer              20110  2 snd_pcm,snd_seq
snd                    58744  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
spidev                  5248  0
evdev                   9450  2
joydev                  9356  0
leds_gpio               2243  0
led_class               3570  1 leds_gpio
spi_bcm2708             4841  0
i2c_bcm2708             3947  0
lsmod Nov 5th update BEFORE python job is run

Code: Select all

Module                  Size  Used by
rtc_ds1307              7767  0
fuse                   72305  3
w1_therm                2945  0
w1_gpio                 2391  0
wire                   21950  2 w1_gpio,w1_therm
8192cu                482155  0
i2c_dev                 5380  16
snd_bcm2835            16213  0
snd_pcm                78007  1 snd_bcm2835
snd_page_alloc          5184  1 snd_pcm
snd_seq                51559  0
snd_seq_device          6501  1 snd_seq
snd_timer              18709  2 snd_pcm,snd_seq
snd                    58083  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
joydev                  8636  0
evdev                   8958  2
leds_gpio               2083  0
led_class               3712  1 leds_gpio
spi_bcm2708             4475  0
i2c_bcm2708             3805  0
lsmod with nov 5th update and the python job run to error

Code: Select all

Traceback (most recent call last):
  File "mod3atod.py", line 33, in <module>
    spi.open(0,0)          # The is on SPI channel 0 (CE0 - aka GPIO8)
IOError: [Errno 2] No such file or directory
pi@developmentpi ~/dev/devdata/kitchenpi/cosm/outside_temp $ lsmod
Module                  Size  Used by
md4                     3386  0
sha256_generic          9131  1
nls_utf8                1247  1
cifs                  203521  2
rtc_ds1307              7767  0
fuse                   72305  3
w1_therm                2945  0
w1_gpio                 2391  0
wire                   21950  2 w1_gpio,w1_therm
8192cu                482155  0
i2c_dev                 5380  16
snd_bcm2835            16213  0
snd_pcm                78007  1 snd_bcm2835
snd_page_alloc          5184  1 snd_pcm
snd_seq                51559  0
snd_seq_device          6501  1 snd_seq
snd_timer              18709  2 snd_pcm,snd_seq
snd                    58083  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
joydev                  8636  0
evdev                   8958  2
leds_gpio               2083  0
led_class               3712  1 leds_gpio
spi_bcm2708             4475  0
i2c_bcm2708             3805  0

Does this help?

Roger

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 2:32 pm

JerryPi wrote:I would like to see this included (is in master and 3.9) as well:
CONFIG_R8712U=m
(Supports USB devices based on the RTL8188SU, RTL8191SU and RTL8192SU chips).
Thank you!
Added with latest update.

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 2:33 pm

Latest firmware has CONFIG_BLK_DEV_SD=y. Any better for USB disk booters?

Corneille
Posts: 2
Joined: Wed Nov 06, 2013 4:44 pm

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 4:48 pm

dom wrote:Latest firmware has CONFIG_BLK_DEV_SD=y. Any better for USB disk booters?
Yes, this works for me.

RobHenry
Posts: 452
Joined: Fri Sep 21, 2012 9:04 pm
Location: UK

Re: Next firmware tree is now 3.10

Wed Nov 06, 2013 6:00 pm

dom wrote:Latest firmware has CONFIG_BLK_DEV_SD=y. Any better for USB disk booters?
That works for me. Thanks.

JerryPi
Posts: 17
Joined: Tue Nov 20, 2012 8:25 am

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 10:27 am

dom wrote:
JerryPi wrote: CONFIG_R8712U=m
Added with latest update.
Thank you, works like a charm.

As for g7ruh's SPI link problem, the only difference I see is in CONFIG_MMC_SPI=m enabled for 3.6.y and not set in 3.10.
Don't grow up, it's a trap!

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 3:09 pm

g7ruh wrote:Does this help?

Roger
Yes, this is the missing bit:

Code: Select all

cn                      4834  1 wire
You need a cn.ko module (userspace connector). Should be in update later today.

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 4:25 pm

Update has been pushed.

This also includes the HDMI audio fixes from next branch, and sdcard corruption fixes (see https://github.com/raspberrypi/linux/is ... t-27763584),
so testing of those would be appreciated.

User avatar
g7ruh
Posts: 67
Joined: Mon Apr 23, 2012 9:49 am
Location: Blackfield UK

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 5:41 pm

dom wrote: so testing of those would be appreciated.
Dom,

have tested with the two latest versions as I discovered you had done another push while I was testing :? Neither work and give the same fault as yesterday. As soon as I revert to 18 Oct version all works fine.

Extract from bcmstat.sh and lsmod below:

Code: Select all

   Other: TEMP_LIMIT=85C
 Version: Nov  7 2013 15:28:26, version 7fc43cae4787c82b467d5707f48021b65cf3d44e (tainted) (release)
vcg path: /usr/bin/vcgencmd
  Codecs: H264 MPG2 VP8 VORBIS MJPG
  Booted: Thu Nov  7 17:12:10 GMT 2013


pi@developmentpi ~/dev/devdata/kitchenpi/cosm/outside_temp $ lsmod
Module                  Size  Used by
rtc_ds1307              7767  0
fuse                   72305  3
w1_therm                2945  0
w1_gpio                 2391  0
wire                   24656  2 w1_gpio,w1_therm
cn                      4568  1 wire
8192cu                482155  0
i2c_dev                 5380  18
snd_bcm2835            16213  0
snd_pcm                78007  1 snd_bcm2835
snd_page_alloc          5184  1 snd_pcm
snd_seq                51559  0
snd_seq_device          6501  1 snd_seq
snd_timer              18709  2 snd_pcm,snd_seq
snd                    58083  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
evdev                   8958  2
joydev                  8636  0
leds_gpio               2083  0
led_class               3712  1 leds_gpio
i2c_bcm2708             3805  0
spi_bcm2708             4475  0



to revert sudo rpi-update 8234d5148aded657760e9ecd622f324d140ae891   for the previous working version

and another update, still does not work

Max Freq:  700Mhz |  250Mhz |  400Mhz
Voltages:       0, 1.20V    |  0, 1.20V
   Other: TEMP_LIMIT=85C
 Version: Nov  7 2013 16:45:47, version 8573b0747a17baf22e97e08254737a13d80430dd (clean) (release)
vcg path: /usr/bin/vcgencmd
  Codecs: H264 MPG2 VP8 VORBIS MJPG
  Booted: Thu Nov  7 17:23:15 GMT 201

Module                  Size  Used by
rtc_ds1307              7767  0
fuse                   72305  3
w1_therm                2945  0
w1_gpio                 2391  0
wire                   24656  2 w1_gpio,w1_therm
cn                      4568  1 wire
8192cu                482155  0
i2c_dev                 5380  16
snd_bcm2835            16213  0
snd_pcm                78007  1 snd_bcm2835
snd_page_alloc          5184  1 snd_pcm
snd_seq                51559  0
snd_seq_device          6501  1 snd_seq
snd_timer              18709  2 snd_pcm,snd_seq
snd                    58083  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
joydev                  8636  0
evdev                   8958  2
leds_gpio               2083  0
led_class               3712  1 leds_gpio
spi_bcm2708             4475  0
i2c_bcm2708             3805  0

then reverted to
"If no errors appeared, your firmware was successfully updated to revision 8234d5148aded657760e9ecd622f324d140ae891"


   Other: TEMP_LIMIT=85C
 Version: Oct 18 2013 16:04:29, version dbda126981820ea5681263042435388dddf65131 (tainted) (release)
vcg path: /usr/bin/vcgencmd
  Codecs: H264 MPG2 VP8 VORBIS MJPG
  Booted: Thu Nov  7 17:32:39 GMT 2013

lsmod before run python (working version)

Module                  Size  Used by
rtc_ds1307              7042  0
fuse                   69438  3
w1_therm                2995  0
w1_gpio                 1633  0
wire                   24973  2 w1_gpio,w1_therm
cn                      4834  1 wire
8192cu                490361  0
i2c_dev                 5594  16
snd_bcm2835            16432  0
snd_pcm                77728  1 snd_bcm2835
snd_page_alloc          5169  1 snd_pcm
snd_seq                53482  0
snd_seq_device          6462  1 snd_seq
snd_timer              20110  2 snd_pcm,snd_seq
snd                    58744  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
spidev                  5248  0
joydev                  9356  0
evdev                   9450  2
leds_gpio               2243  0
led_class               3570  1 leds_gpio
spi_bcm2708             4841  0
i2c_bcm2708             3947  0


after run python (working version)

Module                  Size  Used by
rtc_ds1307              7042  0
fuse                   69438  3
w1_therm                2995  0
w1_gpio                 1633  0
wire                   24973  2 w1_gpio,w1_therm
cn                      4834  1 wire
8192cu                490361  0
i2c_dev                 5594  16
snd_bcm2835            16432  0
snd_pcm                77728  1 snd_bcm2835
snd_page_alloc          5169  1 snd_pcm
snd_seq                53482  0
snd_seq_device          6462  1 snd_seq
snd_timer              20110  2 snd_pcm,snd_seq
snd                    58744  5 snd_bcm2835,snd_timer,snd_pcm,snd_seq,snd_seq_device
spidev                  5248  0
joydev                  9356  0
evdev                   9450  2
leds_gpio               2243  0
led_class               3570  1 leds_gpio
spi_bcm2708             4841  0
i2c_bcm2708             3947  0

and it all works as designed / as previously

I await your further thoughts / suggestions

Roger

J_Rey
Posts: 15
Joined: Sat Oct 26, 2013 10:04 pm
Location: Florida, U.S.A.
Contact: Website Yahoo Messenger AOL

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 5:55 pm

dom wrote:
mcgyver83 wrote:... Please can you provide me the changelog link for this last fw version?
About CMA (dynamic memory splitting)?
...
No change to CMA. It's considered experimental and is not actively supported.
Its been nearly a year now since that was introduced so could you explain what issues have been found with CMA & therefore holding up the use of it? Its worked fine for me so far but I don't overclock either. I would think this should be the default once its reliable.

B.Jay
Posts: 28
Joined: Thu Jul 18, 2013 10:57 pm

Re: Next firmware tree is now 3.10

Thu Nov 07, 2013 10:59 pm

dom wrote:Update has been pushed.

(...) and sdcard corruption (...) so testing of those would be appreciated.
Hi dom.

No go on a 512MB Rev. B (not the one I'm using for USB-boot) with a Transcend 32GB Class 10.

Raising the frequency through raspi-config to 800MHz and rebooting instantly causes issues with the filesystems on the SD card. On first boot with the raised clock the system runs a automated recovery on the FAT32 and ext4 partition, on the next boot it bombs boot with a severe ext4 filesystem inconsistency.

The following cards checked out fine (tested with 800 and 900MHz):
- Kingston 4GB (SD) Class 4
- Kingston 8GB (mSD in SD Adapter) Class 4
- Kingston 32GB (mSD in SD Adapter) Class 4
- ADATA 32GB (mSD in SD Adapter) Class 4
- Samsung 32GB Class 10 (mSD in SD Adapter)

Seems the Transcend cards aren't happy campers when trying to overclock the Pi.

EDIT: As I wrote in another post, I already saw this problem with the Transcend card in the 3.10.xx "next" kernels you provided ahead of the stable release. No big deal if you ask me. I'll keep an eye out for the 'out of the blue' SD card corruption issue, which I can't forcibly reproduce.

Nige C
Posts: 46
Joined: Thu Jul 19, 2012 12:22 am
Location: Hadleigh, Suffolk

Re: Next firmware tree is now 3.10

Fri Nov 08, 2013 1:34 am

After this firmware update my PiFace is no longer detected.
Retired power engineer with a lot of real life experience. Life is an adventure again :)

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Fri Nov 08, 2013 11:24 am

B.Jay wrote:Raising the frequency through raspi-config to 800MHz and rebooting instantly causes issues with the filesystems on the SD card. On first boot with the raised clock the system runs a automated recovery on the FAT32 and ext4 partition, on the next boot it bombs boot with a severe ext4 filesystem inconsistency.
Can you confirm if you applied the config options here:
https://github.com/raspberrypi/linux/is ... t-27763584

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5341
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Next firmware tree is now 3.10

Fri Nov 08, 2013 11:37 am

Nige C wrote:After this firmware update my PiFace is no longer detected.
Did you rpi-update again yesterday afternoon?
I don't know much about PiFace. Anyone know what kernel modules it requires?
All SPI modules from 3.6.y should be present, but perhaps it needs something else.
If you report output of lsmod with PiFace working on older kernel, and after failing on latest kernel, it might indicate what's missing.

Return to “Advanced users”