A major update has been done on TI PPA. Please read the below e-mail
especially for upgrade.
If you do not have installed the TI release PPA, please refer to New
installation procedure.
*New installation procedure:*
1- Add TI OMAP release PPA:
$ sudo su -
# export http_proxy=<your proxy if any>
# export https_proxy=<your proxy if any>
# add-apt-repository ppa:tiomap-dev/release
2- Perform the upgrade:
$ sudo apt-get update
$ sudo apt-get dist-upgrade
$ sudo apt-get install ubuntu-omap4-extras
3- You can re-run a dist-upgrade for safety at the end, then reboot
*Post Installation Important Information:*
* after the installation/update, it is advised to log as default user and
to execute:
# On Panda 4430:
$ /usr/bin/alsaucm -c Panda set _verb HiFi
# On PandaES 4460:
$ /usr/bin/alsaucm -c PandaES set _verb HiFi
* You need to force installation of new bootloaders through command:
$ sudo /usr/sbin/flash-kernel --update-bootloader
* It is advised to use Ubuntu 2D instead of Ubuntu (3D) default UI.
Nevertheless, main compiz issues have been fixed.
* In some cases, you might need to re-install dkms packages: after reboot,
check if omapdrm_pvr module is loaded, if not execute following command
before to reboot:
$ sudo apt-get install --reinstall pvr-omap4-dkms
*New/Fixed in August 2012 release:*
* Kernel bug fixes:
- SmartReflex Class 1.5 enabled with stability patches
- IVAHD Power management (turning off IVAHD in between frames)
- Temperature sensor fix
- Add ARM HW ERRATA 743622, 751472 and 742230, PL310: 769419, OMAP: i659,
i608, i705.
- upstream patches + bug fixes: suspend and wake-up fixes, reduced/fixed
timer interrupt occurrences (twd)
* H264 and MPEG4 video encoder. No camera support yet but Support for
combo use-cases such as: v4l2 usb camera -> preview -> encode -> decode ->
display.
Below is an example of VTC (videoconference) e.g. V4L2 USB webcam +
preview + encode + decode + display, using V4L2, DRI2 and PVR.
gst-launch -v v4l2src ! stridetransform ! queue ! tee name=t \
t. ! queue ! dri2videosink sync=false \
t. ! queue ! ducatih264enc rate-preset=low-delay profile=100 !
h264parse ! ducatih264dec ! queue ! pvrvideosink sync=false
* Decoder issues Fixes.
* Xrandr Rotation Fixes.
*Known Issues:*
* BT is not supported
* Black screen issue on some HW - workaround: boot with screen un-plugged,
plug it after 1 minute. Issue under investigation.
The 'top-level' ubuntu-omap-extras will install the following packages:
- ubuntu-omap-extras-multimedia: another meta package that installs all
the requires packages for MM use case:
* GST, a modified version still compatible with all Ubuntu apps such as
totem
* gst-ducati and libDCE the main components that bring MM h/w
acceleration into the system
- ubuntu-omap-extras-graphics: meta package that pulls in all GFX
components including new X11 driver (binary blob) and SGX kernel module
(source using DKMS to build)
- ubuntu-omap-extras-config: some config scripts to set some permission
issues, and to ensure that the Ducati firmware is loaded at boot time by
upstart.
The system often hangs on booting with black window. After several reboot, it might be OK. But it will hang next time. I didn't know where to file this bug.
On Tuesday, September 4, 2012 8:25:09 PM UTC+8, Boudet, Xavier wrote:
> Hi there,
> A major update has been done on TI PPA. Please read the below e-mail > especially for upgrade. > If you do not have installed the TI release PPA, please refer to New > installation procedure.
> *New installation procedure:* > 1- Add TI OMAP release PPA: > $ sudo su - > # export http_proxy=<your proxy if any> > # export https_proxy=<your proxy if any> > # add-apt-repository ppa:tiomap-dev/release > 2- Perform the upgrade: > $ sudo apt-get update > $ sudo apt-get dist-upgrade > $ sudo apt-get install ubuntu-omap4-extras > 3- You can re-run a dist-upgrade for safety at the end, then reboot
> *Post Installation Important Information:* > * after the installation/update, it is advised to log as default user and > to execute: > # On Panda 4430: > $ /usr/bin/alsaucm -c Panda set _verb HiFi > # On PandaES 4460: > $ /usr/bin/alsaucm -c PandaES set _verb HiFi
> * You need to force installation of new bootloaders through command: > $ sudo /usr/sbin/flash-kernel --update-bootloader
> * It is advised to use Ubuntu 2D instead of Ubuntu (3D) default UI. > Nevertheless, main compiz issues have been fixed.
> * In some cases, you might need to re-install dkms packages: after reboot, > check if omapdrm_pvr module is loaded, if not execute following command > before to reboot: > $ sudo apt-get install --reinstall pvr-omap4-dkms
> *New/Fixed in August 2012 release:* > * Kernel bug fixes:
> - SmartReflex Class 1.5 enabled with stability patches > - IVAHD Power management (turning off IVAHD in between frames) > - Temperature sensor fix > - Add ARM HW ERRATA 743622, 751472 and 742230, PL310: 769419, OMAP: i659, > i608, i705. > - upstream patches + bug fixes: suspend and wake-up fixes, reduced/fixed > timer interrupt occurrences (twd)
> * H264 and MPEG4 video encoder. No camera support yet but Support for > combo use-cases such as: v4l2 usb camera -> preview -> encode -> decode -> > display. > Below is an example of VTC (videoconference) e.g. V4L2 USB webcam + > preview + encode + decode + display, using V4L2, DRI2 and PVR. > gst-launch -v v4l2src ! stridetransform ! queue ! tee name=t \ > t. ! queue ! dri2videosink sync=false \ > t. ! queue ! ducatih264enc rate-preset=low-delay profile=100 ! > h264parse ! ducatih264dec ! queue ! pvrvideosink sync=false > * Decoder issues Fixes. > * Xrandr Rotation Fixes.
> *Known Issues:* > * BT is not supported > * Black screen issue on some HW - workaround: boot with screen un-plugged, > plug it after 1 minute. Issue under investigation.
> The 'top-level' ubuntu-omap-extras will install the following packages: > - ubuntu-omap-extras-multimedia: another meta package that installs all > the requires packages for MM use case: > * GST, a modified version still compatible with all Ubuntu apps such as > totem > * gst-ducati and libDCE the main components that bring MM h/w > acceleration into the system > - ubuntu-omap-extras-graphics: meta package that pulls in all GFX > components including new X11 driver (binary blob) and SGX kernel module > (source using DKMS to build) > - ubuntu-omap-extras-config: some config scripts to set some permission > issues, and to ensure that the Ducati firmware is loaded at boot time by > upstart.
> A major update has been done on TI PPA. Please read the below e-mail > especially for upgrade.
> If you do not have installed the TI release PPA, please refer to New > installation procedure.
All very except that it now seems to have totally broken all my accelerated video (yes, I've rebooted).
For example, if I try to use totem to play video using the hardware acceleration I now get an error "general stream error" and a simple gst-launch -v ducatih264enc gives me
Setting pipeline to PAUSED ...
libdce.c:437: init info: attempting to open X11 connection
libdce.c:515: init error: could not get plugin ioctl base: -22
Setting pipeline to NULL ...
Freeing pipeline ...
Any ideas? Looks like the libdce hardware acceleration has somehow vanished.
One clue: it looks like the ducati is at the wrong version or something. If it do an apt-get install tiomap4-syslink-mm-ducati-firmware then I get:
Reading package lists... Done
Building dependency tree
Reading state information... Done
*The following packages will be REMOVED:
ti-firmware-ipu-dce ubuntu-omap4-extras
ubuntu-omap4-extras-multimedia*
The following NEW packages will be installed:
tiomap4-syslink-mm-ducati-firmware
0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
Need to get 0 B/5,444 kB of archives.
After this operation, 13.1 MB of additional disk space will be used.
Do you want to continue [Y/n]?
On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
> On 04/09/12 13:25, Boudet, Xavier wrote:
> A major update has been done on TI PPA. Please read the below e-mail > especially for upgrade.
> If you do not have installed the TI release PPA, please refer to New > installation procedure.
> All very except that it now seems to have totally broken all my > accelerated video (yes, I've rebooted).
> For example, if I try to use totem to play video using the hardware > acceleration I now get an error “general stream error” and a simple gst-launch > -v ducatih264enc gives me
> Setting pipeline to PAUSED ...
> libdce.c:437: init info: attempting to open X11 connection
> libdce.c:515: init error: could not get plugin ioctl base: -22
> Setting pipeline to NULL ...
> Freeing pipeline ...
> Any ideas? Looks like the libdce hardware acceleration has somehow > vanished.
> One clue: it looks like the ducati is at the wrong version or something. > If it do an apt-get install tiomap4-syslink-mm-ducati-firmware then I get:
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> *The following packages will be REMOVED:
> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia*
> The following NEW packages will be installed:
> tiomap4-syslink-mm-ducati-firmware
> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
> Need to get 0 B/5,444 kB of archives.
> After this operation, 13.1 MB of additional disk space will be used.
> Do you want to continue [Y/n]?
First of all, tiomap4-syslink-mm-ducati-firmware has been replaced by
ti-firmware-ipu-mm.
Fang Bao: can you check if the omapdrm_pvr module is present after boot? If
not, please force re-install with command: sudo apt-get install --reinstall
pvr-omap4-dkms.
Jim: not sure why you are using ducatih264*enc*... Can you just try:
DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file>
video-sink="dri2videosink"
If you are still facing issue, please boot, reproduce the issue you are
facing then execute as root or with sudo: omap-collect-info
Then send me the output file.
On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.bao.wo...@gmail.com> wrote:
> Yeah, this issue is same as my story.
> Further more, I encountered frequent hang in Ubuntu boot step.
> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>> On 04/09/12 13:25, Boudet, Xavier wrote:
>> A major update has been done on TI PPA. Please read the below e-mail
>> especially for upgrade.
>> If you do not have installed the TI release PPA, please refer to New
>> installation procedure.
>> All very except that it now seems to have totally broken all my
>> accelerated video (yes, I've rebooted).
>> For example, if I try to use totem to play video using the hardware
>> acceleration I now get an error “general stream error” and a simple gst-launch
>> -v ducatih264enc gives me
>> Setting pipeline to PAUSED ...
>> libdce.c:437: init info: attempting to open X11 connection
>> libdce.c:515: init error: could not get plugin ioctl base: -22
>> Setting pipeline to NULL ...
>> Freeing pipeline ...
>> Any ideas? Looks like the libdce hardware acceleration has somehow
>> vanished.
>> One clue: it looks like the ducati is at the wrong version or something.
>> If it do an apt-get install tiomap4-syslink-mm-ducati-**firmware then I
>> get:
>> Reading package lists... Done
>> Building dependency tree
>> Reading state information... Done
>> *The following packages will be REMOVED:
>> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia*
>> The following NEW packages will be installed:
>> tiomap4-syslink-mm-ducati-**firmware
>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>> Need to get 0 B/5,444 kB of archives.
>> After this operation, 13.1 MB of additional disk space will be used.
>> Do you want to continue [Y/n]?
On Thu, Sep 6, 2012 at 9:33 AM, Boudet, Xavier <x-bou...@ti.com> wrote:
> Hello,
> Sorry for the late answer.
> First of all, tiomap4-syslink-mm-ducati-firmware has been replaced by
> ti-firmware-ipu-mm.
> Fang Bao: can you check if the omapdrm_pvr module is present after boot?
> If not, please force re-install with command: sudo apt-get install
> --reinstall pvr-omap4-dkms.
> Jim: not sure why you are using ducatih264*enc*... Can you just try:
> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file>
> video-sink="dri2videosink"
> If you are still facing issue, please boot, reproduce the issue you are
> facing then execute as root or with sudo: omap-collect-info
> Then send me the output file.
> Regards,
> *Xavier Boudet *
> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.bao.wo...@gmail.com> wrote:
>> Yeah, this issue is same as my story.
>> Further more, I encountered frequent hang in Ubuntu boot step.
>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>> A major update has been done on TI PPA. Please read the below e-mail
>>> especially for upgrade.
>>> If you do not have installed the TI release PPA, please refer to New
>>> installation procedure.
>>> All very except that it now seems to have totally broken all my
>>> accelerated video (yes, I've rebooted).
>>> For example, if I try to use totem to play video using the hardware
>>> acceleration I now get an error “general stream error” and a simple gst-launch
>>> -v ducatih264enc gives me
>>> Setting pipeline to PAUSED ...
>>> libdce.c:437: init info: attempting to open X11 connection
>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>> Setting pipeline to NULL ...
>>> Freeing pipeline ...
>>> Any ideas? Looks like the libdce hardware acceleration has somehow
>>> vanished.
>>> One clue: it looks like the ducati is at the wrong version or something.
>>> If it do an apt-get install tiomap4-syslink-mm-ducati-**firmware then I
>>> get:
>>> Reading package lists... Done
>>> Building dependency tree
>>> Reading state information... Done
>>> *The following packages will be REMOVED:
>>> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia
>>> *
>>> The following NEW packages will be installed:
>>> tiomap4-syslink-mm-ducati-**firmware
>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>> Need to get 0 B/5,444 kB of archives.
>>> After this operation, 13.1 MB of additional disk space will be used.
>>> Do you want to continue [Y/n]?
On Thursday, September 6, 2012 3:33:42 PM UTC+8, Boudet, Xavier wrote:
> Hello,
> Sorry for the late answer.
> First of all, tiomap4-syslink-mm-ducati-firmware has been replaced by > ti-firmware-ipu-mm.
> Fang Bao: can you check if the omapdrm_pvr module is present after boot? > If not, please force re-install with command: sudo apt-get install > --reinstall pvr-omap4-dkms.
> Jim: not sure why you are using ducatih264*enc*... Can you just try:
> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file> > video-sink="dri2videosink"
> If you are still facing issue, please boot, reproduce the issue you are > facing then execute as root or with sudo: omap-collect-info
> Then send me the output file.
> Regards,
> *Xavier Boudet *
> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.ba...@gmail.com<javascript:>
> > wrote:
>> Yeah, this issue is same as my story.
>> Further more, I encountered frequent hang in Ubuntu boot step.
>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>> A major update has been done on TI PPA. Please read the below e-mail >>> especially for upgrade.
>>> If you do not have installed the TI release PPA, please refer to New >>> installation procedure.
>>> All very except that it now seems to have totally broken all my >>> accelerated video (yes, I've rebooted).
>>> For example, if I try to use totem to play video using the hardware >>> acceleration I now get an error “general stream error” and a simple gst-launch >>> -v ducatih264enc gives me
>>> Setting pipeline to PAUSED ...
>>> libdce.c:437: init info: attempting to open X11 connection
>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>> Setting pipeline to NULL ...
>>> Freeing pipeline ...
>>> Any ideas? Looks like the libdce hardware acceleration has somehow >>> vanished.
>>> One clue: it looks like the ducati is at the wrong version or something. >>> If it do an apt-get install tiomap4-syslink-mm-ducati-**firmware then I >>> get:
>>> Reading package lists... Done
>>> Building dependency tree
>>> Reading state information... Done
>>> *The following packages will be REMOVED:
>>> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia
>>> *
>>> The following NEW packages will be installed:
>>> tiomap4-syslink-mm-ducati-**firmware
>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>> Need to get 0 B/5,444 kB of archives.
>>> After this operation, 13.1 MB of additional disk space will be used.
>>> Do you want to continue [Y/n]?
> You have executed:
> md5sum /lib/firmware/ducati-m3-core0.xem
> Please execute:
> md5sum /lib/firmware/ducati-m3-core0.xem*3*
> You shall get something like that:
> 7a15454a89945bf0ac48ecee92a7a42d /lib/firmware/ducati-m3-core0.xem3
> If not the case please reinstall ti-firmware-ipu-dce.
> On Thu, Sep 6, 2012 at 9:33 AM, Boudet, Xavier <x-bo...@ti.com<javascript:>
> > wrote:
>> Hello,
>> Sorry for the late answer.
>> First of all, tiomap4-syslink-mm-ducati-firmware has been replaced by >> ti-firmware-ipu-mm.
>> Fang Bao: can you check if the omapdrm_pvr module is present after boot? >> If not, please force re-install with command: sudo apt-get install >> --reinstall pvr-omap4-dkms.
>> Jim: not sure why you are using ducatih264*enc*... Can you just try:
>> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file> >> video-sink="dri2videosink"
>> If you are still facing issue, please boot, reproduce the issue you are >> facing then execute as root or with sudo: omap-collect-info
>> Then send me the output file.
>> Regards,
>> *Xavier Boudet *
>> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.ba...@gmail.com<javascript:>
>> > wrote:
>>> Yeah, this issue is same as my story.
>>> Further more, I encountered frequent hang in Ubuntu boot step.
>>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>>> A major update has been done on TI PPA. Please read the below e-mail >>>> especially for upgrade.
>>>> If you do not have installed the TI release PPA, please refer to New >>>> installation procedure.
>>>> All very except that it now seems to have totally broken all my >>>> accelerated video (yes, I've rebooted).
>>>> For example, if I try to use totem to play video using the hardware >>>> acceleration I now get an error “general stream error” and a simple gst-launch >>>> -v ducatih264enc gives me
>>>> Setting pipeline to PAUSED ...
>>>> libdce.c:437: init info: attempting to open X11 connection
>>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>>> Setting pipeline to NULL ...
>>>> Freeing pipeline ...
>>>> Any ideas? Looks like the libdce hardware acceleration has somehow >>>> vanished.
>>>> One clue: it looks like the ducati is at the wrong version or >>>> something. If it do an apt-get install tiomap4-syslink-mm-ducati-**
>>>> firmware then I get:
>>>> Reading package lists... Done
>>>> Building dependency tree
>>>> Reading state information... Done
>>>> *The following packages will be REMOVED:
>>>> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia
>>>> *
>>>> The following NEW packages will be installed:
>>>> tiomap4-syslink-mm-ducati-**firmware
>>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>>> Need to get 0 B/5,444 kB of archives.
>>>> After this operation, 13.1 MB of additional disk space will be used.
>>>> Do you want to continue [Y/n]?
[ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count
25088, align 8)
[ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed
[ 11.482086] cma: dma_alloc_from_contiguous(): memory range at
c0e59000 is busy, retrying
[ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed
[ 11.487792] cma: dma_alloc_from_contiguous(): memory range at
c0e5b000 is busy, retrying
[ 11.493316] alloc_contig_range test_pages_isolated(99800, 9fb00) failed
[ 11.496856] cma: dma_alloc_from_contiguous(): memory range at
c0e5d000 is busy, retrying
[ 11.511535] alloc_contig_range test_pages_isolated(99800, 9fc00) failed
[ 11.514160] cma: dma_alloc_from_contiguous(): memory range at
c0e5f000 is busy, retrying
[ 11.515563] Unable to handle kernel NULL pointer dereference at
virtual address 000000ec
I would be surprised it comes from the bootargs, but indeed you might try.
I answered you the following into the bug:
You need in fact to edit as root the file /boot/boot.script
update the bootargs line as follow:
setenv bootargs ro elevator=noop
mem=1G@0x80000000root=UUID=4dd037b6-24e8-4602-9caa-f13574fb6709 fixrtc
quiet splash
You will then need to execute: sudo flash-kernel
Finally reboot
>> You have executed:
>> md5sum /lib/firmware/ducati-m3-core0.**xem
>> Please execute:
>> md5sum /lib/firmware/ducati-m3-core0.**xem*3*
>> You shall get something like that:
>> 7a15454a89945bf0ac48ecee92a7a4**2d /lib/firmware/ducati-m3-**core0.xem3
>> If not the case please reinstall ti-firmware-ipu-dce.
>> On Thu, Sep 6, 2012 at 9:33 AM, Boudet, Xavier <x-bo...@ti.com> wrote:
>>> Hello,
>>> Sorry for the late answer.
>>> First of all, tiomap4-syslink-mm-**ducati-firmware has been replaced by
>>> ti-firmware-ipu-mm.
>>> Fang Bao: can you check if the omapdrm_pvr module is present after boot?
>>> If not, please force re-install with command: sudo apt-get install
>>> --reinstall pvr-omap4-dkms.
>>> Jim: not sure why you are using ducatih264*enc*... Can you just try:
>>> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file>
>>> video-sink="dri2videosink"
>>> If you are still facing issue, please boot, reproduce the issue you are
>>> facing then execute as root or with sudo: omap-collect-info
>>> Then send me the output file.
>>> Regards,
>>> *Xavier Boudet *
>>> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.ba...@gmail.com> wrote:
>>>> Yeah, this issue is same as my story.
>>>> Further more, I encountered frequent hang in Ubuntu boot step.
>>>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>>>> A major update has been done on TI PPA. Please read the below e-mail
>>>>> especially for upgrade.
>>>>> If you do not have installed the TI release PPA, please refer to New
>>>>> installation procedure.
>>>>> All very except that it now seems to have totally broken all my
>>>>> accelerated video (yes, I've rebooted).
>>>>> For example, if I try to use totem to play video using the hardware
>>>>> acceleration I now get an error “general stream error” and a simple gst-launch
>>>>> -v ducatih264enc gives me
>>>>> Setting pipeline to PAUSED ...
>>>>> libdce.c:437: init info: attempting to open X11 connection
>>>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>>>> Setting pipeline to NULL ...
>>>>> Freeing pipeline ...
>>>>> Any ideas? Looks like the libdce hardware acceleration has somehow
>>>>> vanished.
>>>>> One clue: it looks like the ducati is at the wrong version or
>>>>> something. If it do an apt-get install tiomap4-syslink-mm-ducati-**
>>>>> firm**ware then I get:
>>>>> Reading package lists... Done
>>>>> Building dependency tree
>>>>> Reading state information... Done
>>>>> *The following packages will be REMOVED:
>>>>> ti-firmware-ipu-dce ubuntu-omap4-extras
>>>>> ubuntu-omap4-extras-multimedia*
>>>>> The following NEW packages will be installed:
>>>>> tiomap4-syslink-mm-ducati-**firm**ware
>>>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>>>> Need to get 0 B/5,444 kB of archives.
>>>>> After this operation, 13.1 MB of additional disk space will be used.
>>>>> Do you want to continue [Y/n]?
Thanks Boudet. Following your instructions, the new kernel command line has been flashed into boot image.
After reboot, I find the correct kernel command in dmesg log.
But when I run totem to play H264 stream, there is a new error:
$ totem XXXXXXX.MP4 libdce.c:437: init info: attempting to open X11 connection
** Message: Error: Could not decode stream.
gstducatividdec.c(1081): gst_ducati_viddec_chain (): /GstPlayBin2:play/GstURIDecodeBin:uridecodebin0/GstDecodeBin2:decodebin20/G stDucatiH264Dec:ducatih264dec0:
process returned error: -1 08008000
When I tried again, there is no error printed. But totem can't play stream also.
> [ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count 25088, align 8)
> [ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed
> [ 11.482086] cma: dma_alloc_from_contiguous(): memory range at c0e59000 is busy, retrying
> [ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed
> [ 11.487792] cma: dma_alloc_from_contiguous(): memory range at c0e5b000 is busy, retrying
> [ 11.493316] alloc_contig_range test_pages_isolated(99800, 9fb00) failed
> [ 11.496856] cma: dma_alloc_from_contiguous(): memory range at c0e5d000 is busy, retrying
> [ 11.511535] alloc_contig_range test_pages_isolated(99800, 9fc00) failed
> [ 11.514160] cma: dma_alloc_from_contiguous(): memory range at c0e5f000 is busy, retrying
> [ 11.515563] Unable to handle kernel NULL pointer dereference at virtual address 000000ec
> I would be surprised it comes from the bootargs, but indeed you might try.
> I answered you the following into the bug:
> You need in fact to edit as root the file /boot/boot.script
> update the bootargs line as follow:
> setenv bootargs ro elevator=noop mem=1G@0x80000000 > root=UUID=4dd037b6-24e8-4602-9caa-f13574fb6709 fixrtc quiet splash
> You will then need to execute: sudo flash-kernel
> Finally reboot
> *Xavier Boudet*
> On Thu, Sep 6, 2012 at 11:39 AM, fang bao <fang.ba...@gmail.com<javascript:>
> > wrote:
>> I have confirmed. The md5 is correct.
>> # md5sum /lib/firmware/ducati*
>> 7a15454a89945bf**0ac48ecee92a7a4**2d /lib/firmware/**ducati-**m3-core0.**
>> xem3
>> On Thursday, September 6, 2012 5:17:37 PM UTC+8, Boudet, Xavier wrote:
>>> You have executed:
>>> md5sum /lib/firmware/ducati-m3-core0.**xem
>>> Please execute:
>>> md5sum /lib/firmware/ducati-m3-core0.**xem*3*
>>> You shall get something like that:
>>> 7a15454a89945bf0ac48ecee92a7a4**2d /lib/firmware/ducati-m3-**core0.xem3
>>> If not the case please reinstall ti-firmware-ipu-dce.
>>> On Thu, Sep 6, 2012 at 9:33 AM, Boudet, Xavier <x-bo...@ti.com> wrote:
>>>> Hello,
>>>> Sorry for the late answer.
>>>> First of all, tiomap4-syslink-mm-**ducati-firmware has been replaced >>>> by ti-firmware-ipu-mm.
>>>> Fang Bao: can you check if the omapdrm_pvr module is present after >>>> boot? If not, please force re-install with command: sudo apt-get >>>> install --reinstall pvr-omap4-dkms.
>>>> Jim: not sure why you are using ducatih264*enc*... Can you just try:
>>>> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file> >>>> video-sink="dri2videosink"
>>>> If you are still facing issue, please boot, reproduce the issue you are >>>> facing then execute as root or with sudo: omap-collect-info
>>>> Then send me the output file.
>>>> Regards,
>>>> *Xavier Boudet *
>>>> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.ba...@gmail.com> wrote:
>>>>> Yeah, this issue is same as my story.
>>>>> Further more, I encountered frequent hang in Ubuntu boot step.
>>>>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>>>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>>>>> A major update has been done on TI PPA. Please read the below e-mail >>>>>> especially for upgrade.
>>>>>> If you do not have installed the TI release PPA, please refer to New >>>>>> installation procedure.
>>>>>> All very except that it now seems to have totally broken all my >>>>>> accelerated video (yes, I've rebooted).
>>>>>> For example, if I try to use totem to play video using the hardware >>>>>> acceleration I now get an error “general stream error” and a simple gst-launch >>>>>> -v ducatih264enc gives me
>>>>>> Setting pipeline to PAUSED ...
>>>>>> libdce.c:437: init info: attempting to open X11 connection
>>>>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>>>>> Setting pipeline to NULL ...
>>>>>> Freeing pipeline ...
>>>>>> Any ideas? Looks like the libdce hardware acceleration has somehow >>>>>> vanished.
>>>>>> One clue: it looks like the ducati is at the wrong version or >>>>>> something. If it do an apt-get install tiomap4-syslink-mm-ducati-**
>>>>>> firm**ware then I get:
>>>>>> Reading package lists... Done
>>>>>> Building dependency tree
>>>>>> Reading state information... Done
>>>>>> *The following packages will be REMOVED:
>>>>>> ti-firmware-ipu-dce ubuntu-omap4-extras >>>>>> ubuntu-omap4-extras-multimedia*
>>>>>> The following NEW packages will be installed:
>>>>>> tiomap4-syslink-mm-ducati-**firm**ware
>>>>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>>>>> Need to get 0 B/5,444 kB of archives.
>>>>>> After this operation, 13.1 MB of additional disk space will be used.
>>>>>> Do you want to continue [Y/n]?
Do you still have the same messages in dmesg?
[ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count 25088,
align 8)
[ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed
[ 11.482086] cma: dma_alloc_from_contiguous(): memory range at c0e59000
is busy, retrying
[ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed
...
On Thu, Sep 6, 2012 at 12:12 PM, fang bao <fang.bao.wo...@gmail.com> wrote:
> Thanks Boudet. Following your instructions, the new kernel command line
> has been flashed into boot image.
> After reboot, I find the correct kernel command in dmesg log.
> But when I run totem to play H264 stream, there is a new error:
> $ totem XXXXXXX.MP4
> libdce.c:437: init info: attempting to open X11 connection
> ** Message: Error: Could not decode stream.
> gstducatividdec.c(1081): gst_ducati_viddec_chain ():
> /GstPlayBin2:play/GstURIDecodeBin:uridecodebin0/GstDecodeBin2:decodebin20/G stDucatiH264Dec:ducatih264dec0:
> process returned error: -1 08008000
> When I tried again, there is no error printed. But totem can't play stream
> also.
> So the HW MM codec doesn't work.
> On Thursday, September 6, 2012 5:51:19 PM UTC+8, Boudet, Xavier wrote:
>> [ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count 25088, align 8)
>> [ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed
>> [ 11.482086] cma: dma_alloc_from_contiguous(): memory range at c0e59000 is busy, retrying
>> [ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed
>> [ 11.487792] cma: dma_alloc_from_contiguous(): memory range at c0e5b000 is busy, retrying
>> [ 11.493316] alloc_contig_range test_pages_isolated(99800, 9fb00) failed
>> [ 11.496856] cma: dma_alloc_from_contiguous(): memory range at c0e5d000 is busy, retrying
>> [ 11.511535] alloc_contig_range test_pages_isolated(99800, 9fc00) failed
>> [ 11.514160] cma: dma_alloc_from_contiguous(): memory range at c0e5f000 is busy, retrying
>> [ 11.515563] Unable to handle kernel NULL pointer dereference at virtual address 000000ec
>> I would be surprised it comes from the bootargs, but indeed you might try.
>> I answered you the following into the bug:
>> You need in fact to edit as root the file /boot/boot.script
>> update the bootargs line as follow:
>> setenv bootargs ro elevator=noop mem=1G@0x80000000root=UUID=4dd037b6-24e8-4602-
>> **9caa-f13574fb6709 fixrtc quiet splash
>> You will then need to execute: sudo flash-kernel
>> Finally reboot
>> *Xavier Boudet*
>> On Thu, Sep 6, 2012 at 11:39 AM, fang bao <fang.ba...@gmail.com> wrote:
>>> I have confirmed. The md5 is correct.
>>> # md5sum /lib/firmware/ducati*
>>> 7a15454a89945bf**0ac48ecee92a7a4****2d /lib/firmware/**ducati-**
>>> m3-core0.****xem3
>>> On Thursday, September 6, 2012 5:17:37 PM UTC+8, Boudet, Xavier wrote:
>>>> You have executed:
>>>> md5sum /lib/firmware/ducati-m3-core0.****xem
>>>> Please execute:
>>>> md5sum /lib/firmware/ducati-m3-core0.****xem*3*
>>>> You shall get something like that:
>>>> 7a15454a89945bf0ac48ecee92a7a4****2d /lib/firmware/ducati-m3-**core0**
>>>> .xem3
>>>> If not the case please reinstall ti-firmware-ipu-dce.
>>>> On Thu, Sep 6, 2012 at 9:33 AM, Boudet, Xavier <x-bo...@ti.com> wrote:
>>>>> Hello,
>>>>> Sorry for the late answer.
>>>>> First of all, tiomap4-syslink-mm-**ducati**-firmware has been
>>>>> replaced by ti-firmware-ipu-mm.
>>>>> Fang Bao: can you check if the omapdrm_pvr module is present after
>>>>> boot? If not, please force re-install with command: sudo apt-get
>>>>> install --reinstall pvr-omap4-dkms.
>>>>> Jim: not sure why you are using ducatih264*enc*... Can you just try:
>>>>> DISPLAY=:0 gst-launch-0.10 playbin2 uri=file://<your file>
>>>>> video-sink="dri2videosink"
>>>>> If you are still facing issue, please boot, reproduce the issue you
>>>>> are facing then execute as root or with sudo: omap-collect-info
>>>>> Then send me the output file.
>>>>> Regards,
>>>>> *Xavier Boudet *
>>>>> On Thu, Sep 6, 2012 at 8:41 AM, fang bao <fang.ba...@gmail.com> wrote:
>>>>>> Yeah, this issue is same as my story.
>>>>>> Further more, I encountered frequent hang in Ubuntu boot step.
>>>>>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>>>>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>>>>>> A major update has been done on TI PPA. Please read the below e-mail
>>>>>>> especially for upgrade.
>>>>>>> If you do not have installed the TI release PPA, please refer to New
>>>>>>> installation procedure.
>>>>>>> All very except that it now seems to have totally broken all my
>>>>>>> accelerated video (yes, I've rebooted).
>>>>>>> For example, if I try to use totem to play video using the hardware
>>>>>>> acceleration I now get an error “general stream error” and a simple gst-launch
>>>>>>> -v ducatih264enc gives me
>>>>>>> Setting pipeline to PAUSED ...
>>>>>>> libdce.c:437: init info: attempting to open X11 connection
>>>>>>> libdce.c:515: init error: could not get plugin ioctl base: -22
>>>>>>> Setting pipeline to NULL ...
>>>>>>> Freeing pipeline ...
>>>>>>> Any ideas? Looks like the libdce hardware acceleration has somehow
>>>>>>> vanished.
>>>>>>> One clue: it looks like the ducati is at the wrong version or
>>>>>>> something. If it do an apt-get install tiomap4-syslink-mm-ducati-**
>>>>>>> firm****ware then I get:
>>>>>>> Reading package lists... Done
>>>>>>> Building dependency tree
>>>>>>> Reading state information... Done
>>>>>>> *The following packages will be REMOVED:
>>>>>>> ti-firmware-ipu-dce ubuntu-omap4-extras
>>>>>>> ubuntu-omap4-extras-multimedia*
>>>>>>> The following NEW packages will be installed:
>>>>>>> tiomap4-syslink-mm-ducati-**firm****ware
>>>>>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>>>>>> Need to get 0 B/5,444 kB of archives.
>>>>>>> After this operation, 13.1 MB of additional disk space will be used.
>>>>>>> Do you want to continue [Y/n]?
On Thursday, September 6, 2012 6:58:03 PM UTC+8, Boudet, Xavier wrote:
> Do you still have the same messages in dmesg? > [ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count 25088, > align 8) > [ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed > [ 11.482086] cma: dma_alloc_from_contiguous(): memory range at c0e59000 > is busy, retrying > [ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed > ...
On Thu, Sep 6, 2012 at 1:21 PM, fang bao <fang.bao.wo...@gmail.com> wrote:
> No, there is no related failed mssage in dmesg. All related messages are
> listed below:
> On Thursday, September 6, 2012 6:58:03 PM UTC+8, Boudet, Xavier wrote:
>> Do you still have the same messages in dmesg?
>> [ 11.473022] cma: dma_alloc_from_contiguous(cma ed10bdc0, count 25088,
>> align 8)
>> [ 11.479644] alloc_contig_range test_pages_isolated(99700, 9f900) failed
>> [ 11.482086] cma: dma_alloc_from_contiguous(): memory range at c0e59000
>> is busy, retrying
>> [ 11.485717] alloc_contig_range test_pages_isolated(99800, 9fa00) failed
>> ...
Yes, I too see the black screen on boot up. It's not actually a hang but some sort of failure to start the X server properly.
Oddly, the workaround is to switch to the control using Control-Alt-F1 and then back to X with Control-Alt-F7.
I have no explanation for what this problem is or why the workaround works but it seems to.
I didn't mention it in my original e-mail to avoid complicating the discussion.
In related news I logged into my pandaboard remotely (I'm not at the same location right now), modified the /boot/boot.script and it now seems to work a little better. I'll report in later when I can log onto the screen and try the other tests.
Regards,
Jim.
p.s. Thanks to everyone for the help in diagnosing this problem.
> Yeah, this issue is same as my story.
> Further more, I encountered frequent hang in Ubuntu boot step.
> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
> On 04/09/12 13:25, Boudet, Xavier wrote:
>> A major update has been done on TI PPA. Please read the below
>> e-mail especially for upgrade.
>> If you do not have installed the TI release PPA, please refer to
>> New installation procedure.
> All very except that it now seems to have totally broken all my
> accelerated video (yes, I've rebooted).
> For example, if I try to use totem to play video using the
> hardware acceleration I now get an error “general stream error”
> and a simple gst-launch -v ducatih264enc gives me
> Setting pipeline to PAUSED ...
> libdce.c:437: init info: attempting to open X11 connection
> libdce.c:515: init error: could not get plugin ioctl
> base: -22
> Setting pipeline to NULL ...
> Freeing pipeline ...
> Any ideas? Looks like the libdce hardware acceleration has somehow
> vanished.
> One clue: it looks like the ducati is at the wrong version or
> something. If it do an apt-get install
> tiomap4-syslink-mm-ducati-firmware then I get:
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> *The following packages will be REMOVED:
> ti-firmware-ipu-dce ubuntu-omap4-extras
> ubuntu-omap4-extras-multimedia*
> The following NEW packages will be installed:
> tiomap4-syslink-mm-ducati-firmware
> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
> Need to get 0 B/5,444 kB of archives.
> After this operation, 13.1 MB of additional disk space will be
> used.
> Do you want to continue [Y/n]?
On Thu, Sep 6, 2012 at 3:46 PM, Jim Darby <uberscuba...@gmail.com> wrote:
> Just a very simple update.
> Yes, I too see the black screen on boot up. It's not actually a hang but
> some sort of failure to start the X server properly.
> Oddly, the workaround is to switch to the control using Control-Alt-F1 and
> then back to X with Control-Alt-F7.
> I have no explanation for what this problem is or why the workaround works
> but it seems to.
> I didn't mention it in my original e-mail to avoid complicating the
> discussion.
> In related news I logged into my pandaboard remotely (I'm not at the same
> location right now), modified the /boot/boot.script and it now seems to
> work a little better. I'll report in later when I can log onto the screen
> and try the other tests.
> Regards,
> Jim.
> p.s. Thanks to everyone for the help in diagnosing this problem.
> On 06/09/12 07:41, fang bao wrote:
> Yeah, this issue is same as my story.
> Further more, I encountered frequent hang in Ubuntu boot step.
> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>> On 04/09/12 13:25, Boudet, Xavier wrote:
>> A major update has been done on TI PPA. Please read the below e-mail
>> especially for upgrade.
>> If you do not have installed the TI release PPA, please refer to New
>> installation procedure.
>> All very except that it now seems to have totally broken all my
>> accelerated video (yes, I've rebooted).
>> For example, if I try to use totem to play video using the hardware
>> acceleration I now get an error “general stream error” and a simple gst-launch
>> -v ducatih264enc gives me
>> Setting pipeline to PAUSED ...
>> libdce.c:437: init info: attempting to open X11 connection
>> libdce.c:515: init error: could not get plugin ioctl base: -22
>> Setting pipeline to NULL ...
>> Freeing pipeline ...
>> Any ideas? Looks like the libdce hardware acceleration has somehow
>> vanished.
>> One clue: it looks like the ducati is at the wrong version or something.
>> If it do an apt-get install tiomap4-syslink-mm-ducati-**firmware then I
>> get:
>> Reading package lists... Done
>> Building dependency tree
>> Reading state information... Done
>> *The following packages will be REMOVED:
>> ti-firmware-ipu-dce ubuntu-omap4-extras ubuntu-omap4-extras-multimedia*
>> The following NEW packages will be installed:
>> tiomap4-syslink-mm-ducati-**firmware
>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>> Need to get 0 B/5,444 kB of archives.
>> After this operation, 13.1 MB of additional disk space will be used.
>> Do you want to continue [Y/n]?
Bingo! After the modifications to /boot/boot.script it now works perfectly. Many thanks for your prompt help in fixing this.
Can I just ask what on earth all the mem= runes were? It looks like it was trying to place RAM all over the machines address space. The latest version looks like it puts all the RAM in one place half way up the physical address space. This seems a little curious, is there any specific reason for it?
For that reason why was it split before and what happened to the vram parameter (or whatever it was called)? I'[m very interested in the PandaBoard's architecture.
Also, not that I'm being ungrateful, but we still have the minor irritation of the black login screen.
Finally I was looking at the options for ducatih264enc. It would appear that it can handle constant bit rate and various levels of the standard but there doesn't seem to be an option to allow variable bit rate with a fixed quality. Am I missing something here?
> Thanks Jim,
> If you are still seeing issue, do not hesitate to collect data though > : sudo omap-collect-info
> On Thu, Sep 6, 2012 at 3:46 PM, Jim Darby <uberscuba...@gmail.com > <mailto:uberscuba...@gmail.com>> wrote:
> Just a very simple update.
> Yes, I too see the black screen on boot up. It's not actually a
> hang but some sort of failure to start the X server properly.
> Oddly, the workaround is to switch to the control using
> Control-Alt-F1 and then back to X with Control-Alt-F7.
> I have no explanation for what this problem is or why the
> workaround works but it seems to.
> I didn't mention it in my original e-mail to avoid complicating
> the discussion.
> In related news I logged into my pandaboard remotely (I'm not at
> the same location right now), modified the /boot/boot.script and
> it now seems to work a little better. I'll report in later when I
> can log onto the screen and try the other tests.
> Regards,
> Jim.
> p.s. Thanks to everyone for the help in diagnosing this problem.
> On 06/09/12 07:41, fang bao wrote:
>> Yeah, this issue is same as my story.
>> Further more, I encountered frequent hang in Ubuntu boot step.
>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>> A major update has been done on TI PPA. Please read the
>>> below e-mail especially for upgrade.
>>> If you do not have installed the TI release PPA, please
>>> refer to New installation procedure.
>> All very except that it now seems to have totally broken all
>> my accelerated video (yes, I've rebooted).
>> For example, if I try to use totem to play video using the
>> hardware acceleration I now get an error �general stream
>> error� and a simple gst-launch -v ducatih264enc gives me
>> Setting pipeline to PAUSED ...
>> libdce.c:437: init info: attempting to open X11
>> connection
>> libdce.c:515: init error: could not get plugin ioctl
>> base: -22
>> Setting pipeline to NULL ...
>> Freeing pipeline ...
>> Any ideas? Looks like the libdce hardware acceleration has
>> somehow vanished.
>> One clue: it looks like the ducati is at the wrong version or
>> something. If it do an apt-get install
>> tiomap4-syslink-mm-ducati-firmware then I get:
>> Reading package lists... Done
>> Building dependency tree
>> Reading state information... Done
>> *The following packages will be REMOVED:
>> ti-firmware-ipu-dce ubuntu-omap4-extras
>> ubuntu-omap4-extras-multimedia*
>> The following NEW packages will be installed:
>> tiomap4-syslink-mm-ducati-firmware
>> 0 upgraded, 1 newly installed, 3 to remove and 0 not
>> upgraded.
>> Need to get 0 B/5,444 kB of archives.
>> After this operation, 13.1 MB of additional disk space
>> will be used.
>> Do you want to continue [Y/n]?
On Friday, September 7, 2012 3:03:41 AM UTC+8, JimD wrote:
> Bingo! After the modifications to /boot/boot.script it now works > perfectly. Many thanks for your prompt help in fixing this.
> Can I just ask what on earth all the mem= runes were? It looks like it was > trying to place RAM all over the machines address space. The latest version > looks like it puts all the RAM in one place half way up the physical > address space. This seems a little curious, is there any specific reason > for it?
> For that reason why was it split before and what happened to the vram > parameter (or whatever it was called)? I'[m very interested in the > PandaBoard's architecture.
> Also, not that I'm being ungrateful, but we still have the minor > irritation of the black login screen.
> Finally I was looking at the options for ducatih264enc. It would appear > that it can handle constant bit rate and various levels of the standard but > there doesn't seem to be an option to allow variable bit rate with a fixed > quality. Am I missing something here?
> Once again, many thanks for your help.
> Jim.
> On 06/09/12 15:19, Boudet, Xavier wrote:
> Thanks Jim, > If you are still seeing issue, do not hesitate to�collect�data though > : sudo omap-collect-info
> On Thu, Sep 6, 2012 at 3:46 PM, Jim Darby <ubersc...@gmail.com<javascript:>
> > wrote:
>> Just a very simple update.
>> Yes, I too see the black screen on boot up. It's not actually a hang but >> some sort of failure to start the X server properly.
>> Oddly, the workaround is to switch to the control using Control-Alt-F1 >> and then back to X with Control-Alt-F7.
>> I have no explanation for what this problem is or why the workaround >> works but it seems to.
>> I didn't mention it in my original e-mail to avoid complicating the >> discussion.
>> In related news I logged into my pandaboard remotely (I'm not at the same >> location right now), modified the /boot/boot.script and it now seems to >> work a little better. I'll report in later when I can log onto the screen >> and try the other tests.
>> Regards,
>> Jim.
>> p.s. Thanks to everyone for the help in diagnosing this problem.
>> On 06/09/12 07:41, fang bao wrote:
>> Yeah, this issue is same as my story.
>> Further more, I encountered frequent hang in Ubuntu boot step.
>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>> A major update has been done on TI PPA. Please read the below e-mail >>> especially for upgrade.
>>> If you do not have installed the TI release PPA, please refer to New >>> installation procedure.
>>> All very except that it now seems to have totally broken all my >>> accelerated video (yes, I've rebooted).
>>> For example, if I try to use totem to play video using the hardware >>> acceleration I now get an error �general stream error� and a simple gst-launch >>> -v ducatih264enc gives me
>>> Setting pipeline to PAUSED ...
>>> libdce.c:437:�� init��� info: attempting to open X11 connection
>>> libdce.c:515:�� init��� error: could not get plugin ioctl >>> base: -22
>>> Setting pipeline to NULL ...
>>> Freeing pipeline ...
>>> Any ideas? Looks like the libdce hardware acceleration has somehow >>> vanished.
>>> One clue: it looks like the ducati is at the wrong version or something. >>> If it do an apt-get install tiomap4-syslink-mm-ducati-firmware then I >>> get:
>>> Reading package lists... Done
>>> Building dependency tree
>>> Reading state information... Done
>>> *The following packages will be REMOVED:
>>> � ti-firmware-ipu-dce ubuntu-omap4-extras >>> ubuntu-omap4-extras-multimedia*
>>> The following NEW packages will be installed:
>>> � tiomap4-syslink-mm-ducati-firmware
>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not upgraded.
>>> Need to get 0 B/5,444 kB of archives.
>>> After this operation, 13.1 MB of additional disk space will be used.
>>> Do you want to continue [Y/n]?
I'm not sure how many of the filters are needed, especially videorate and ffmpegcolorspace.
Hope this helps,
By the way, you need to run this in a console under X for it to work properly. Otherwise it can't get hold of the DRI infrastructure that controls the DSP for encoding.
> On Friday, September 7, 2012 3:03:41 AM UTC+8, JimD wrote:
> Bingo! After the modifications to /boot/boot.script it now works
> perfectly. Many thanks for your prompt help in fixing this.
> Can I just ask what on earth all the mem= runes were? It looks
> like it was trying to place RAM all over the machines address
> space. The latest version looks like it puts all the RAM in one
> place half way up the physical address space. This seems a little
> curious, is there any specific reason for it?
> For that reason why was it split before and what happened to the
> vram parameter (or whatever it was called)? I'[m very interested
> in the PandaBoard's architecture.
> Also, not that I'm being ungrateful, but we still have the minor
> irritation of the black login screen.
> Finally I was looking at the options for ducatih264enc. It would
> appear that it can handle constant bit rate and various levels of
> the standard but there doesn't seem to be an option to allow
> variable bit rate with a fixed quality. Am I missing something here?
> Once again, many thanks for your help.
> Jim.
> On 06/09/12 15:19, Boudet, Xavier wrote:
>> Thanks Jim,
>> If you are still seeing issue, do not hesitate
>> to�collect�data though : sudo omap-collect-info
>> On Thu, Sep 6, 2012 at 3:46 PM, Jim Darby <ubersc...@gmail.com
>> <javascript:>> wrote:
>> Just a very simple update.
>> Yes, I too see the black screen on boot up. It's not actually
>> a hang but some sort of failure to start the X server properly.
>> Oddly, the workaround is to switch to the control using
>> Control-Alt-F1 and then back to X with Control-Alt-F7.
>> I have no explanation for what this problem is or why the
>> workaround works but it seems to.
>> I didn't mention it in my original e-mail to avoid
>> complicating the discussion.
>> In related news I logged into my pandaboard remotely (I'm not
>> at the same location right now), modified the
>> /boot/boot.script and it now seems to work a little better.
>> I'll report in later when I can log onto the screen and try
>> the other tests.
>> Regards,
>> Jim.
>> p.s. Thanks to everyone for the help in diagnosing this problem.
>> On 06/09/12 07:41, fang bao wrote:
>>> Yeah, this issue is same as my story.
>>> Further more, I encountered frequent hang in Ubuntu boot step.
>>> On Wednesday, September 5, 2012 5:28:57 AM UTC+8, JimD wrote:
>>> On 04/09/12 13:25, Boudet, Xavier wrote:
>>>> A major update has been done on TI PPA. Please read the
>>>> below e-mail especially for upgrade.
>>>> If you do not have installed the TI release PPA, please
>>>> refer to New installation procedure.
>>> All very except that it now seems to have totally broken
>>> all my accelerated video (yes, I've rebooted).
>>> For example, if I try to use totem to play video using
>>> the hardware acceleration I now get an error �general
>>> stream error� and a simple gst-launch -v ducatih264enc
>>> gives me
>>> Setting pipeline to PAUSED ...
>>> libdce.c:437:�� init��� info: attempting
>>> to open X11 connection
>>> libdce.c:515:�� init��� error: could not
>>> get plugin ioctl base: -22
>>> Setting pipeline to NULL ...
>>> Freeing pipeline ...
>>> Any ideas? Looks like the libdce hardware acceleration
>>> has somehow vanished.
>>> One clue: it looks like the ducati is at the wrong
>>> version or something. If it do an apt-get install
>>> tiomap4-syslink-mm-ducati-firmware then I get:
>>> Reading package lists... Done
>>> Building dependency tree
>>> Reading state information... Done
>>> *The following packages will be REMOVED:
>>> � ti-firmware-ipu-dce ubuntu-omap4-extras
>>> ubuntu-omap4-extras-multimedia*
>>> The following NEW packages will be installed:
>>> � tiomap4-syslink-mm-ducati-firmware
>>> 0 upgraded, 1 newly installed, 3 to remove and 0 not
>>> upgraded.
>>> Need to get 0 B/5,444 kB of archives.
>>> After this operation, 13.1 MB of additional disk
>>> space will be used.
>>> Do you want to continue [Y/n]?
On Thu, Sep 6, 2012 at 9:03 PM, Jim Darby <uberscuba...@gmail.com> wrote:
> Bingo! After the modifications to /boot/boot.script it now works perfectly.
> Many thanks for your prompt help in fixing this.
cool!
> Can I just ask what on earth all the mem= runes were? It looks like it was
> trying to place RAM all over the machines address space. The latest version
> looks like it puts all the RAM in one place half way up the physical address
> space. This seems a little curious, is there any specific reason for it?
well, since we migrated to 12.04 we have changed how the memory
allocation is done for h/w accelerator (IVAHD/Ducati). It used to be
that we were making explicit 'memory holes' in the linux memory map by
using specific 'mem=' command in the bootargs. Generally we would
reserve (carevout) ~50Mb, right before the 512Mb (so pretty much in
the middle of the 1G).
that's why you needed to have the exact same bootargs that we
expected. this address was hard coded in the kernel, and ducati
firmware.
with the migration to the new syslink, called rpmsg, we have changed
the way the memory is allocated, and we know rely on the (now merged
in mainline) CMA feature. I think your problems were because CMA
wasn't able to allocate the memory because you had created these
artifical holes in the memory map.
> For that reason why was it split before and what happened to the vram
> parameter (or whatever it was called)? I'[m very interested in the
> PandaBoard's architecture.
the vram was a parameter used by the previous video driver, and which
is not used anymore with the new omapdrm driver.
> Also, not that I'm being ungrateful, but we still have the minor irritation
> of the black login screen.
> Finally I was looking at the options for ducatih264enc. It would appear that
> it can handle constant bit rate and various levels of the standard but there
> doesn't seem to be an option to allow variable bit rate with a fixed
> quality. Am I missing something here?
i would have to check that, as i don't know. perhaps rob or Olivier would know.
I've also noticed another curious issue with the new kernel.
I've put the console=ttyO2,115200n8 at the end of my setenv in my /boot/boot.script and most of the kernel output appears on the serial port. At least for a short period after boot.
However, after a while many of the lines fail to appear. Note that there are no half lines: the line is either entirely there or not there at all. Mostly the latter.
Any idea what's going on here?
I'm not complaining, I just want to be able to help us produce a better kernel.
Regards,
Jim.
p.s. The new kernel does seem a lot more stable. The reason that I added the setenv option was to capture debugging when the kernel panicked. So now it's not panicking nearly as much but when it does (once) I can't capture the debug output. By the way, what provoked the only panic was running totem on a 720p film.
I thought the PandaBoard was supposed to be able to do 1080? Even the 720 film was running at over 100% CPU utilisation (yes, I know it's dual core). It is using the hardware acceleration though. Without it the results are /terrible/!
p.p.s. The initial black login screen problem is still there.
On Fri, Sep 7, 2012 at 4:58 PM, Jim Darby <uberscuba...@gmail.com> wrote:
> p.s. The new kernel does seem a lot more stable. The reason that I added the
> setenv option was to capture debugging when the kernel panicked. So now it's
> not panicking nearly as much but when it does (once) I can't capture the
> debug output. By the way, what provoked the only panic was running totem on
> a 720p film.
> I thought the PandaBoard was supposed to be able to do 1080? Even the 720
> film was running at over 100% CPU utilisation (yes, I know it's dual core).
> It is using the hardware acceleration though. Without it the results are
> terrible!
a 1080p movie playback should be less than 20% CPU load if you are
using IVAHD h/w accelerator and GPU to render... if you don't get
that, there is a problem.
note that only MPEG4, H264, MPEG2 and VC1 are h/w accelerated, not VP*
> a 1080p movie playback should be less than 20% CPU load if you are > using IVAHD h/w accelerator and GPU to render... if you don't get > that, there is a problem. note that only MPEG4, H264, MPEG2 and VC1 > are h/w accelerated, not VP*
OK, it was a H.264 encoded 720p film so it should have been better. I've just checked and it seems pretty consistent. Also with a 576p input I seem to be getting around 75% CPU. I've check and GStreamer is opening /dev/dri/card0.
Also I think I have some sort of CPU frequency changing going on. The BogoMIPS in /proc/cpuinfo change depending on CPU loading so that may be affecting it. However, I don't seem to have any of the cpufreq modules loaded....
Could you suggest some ways of diagnosing this more? Specifically, how could I check if the hardware decoding is actually working?
On Fri, Sep 7, 2012 at 5:31 PM, Jim Darby <uberscuba...@gmail.com> wrote:
> OK, it was a H.264 encoded 720p film so it should have been better. I've
> just checked and it seems pretty consistent. Also with a 576p input I seem
> to be getting around 75% CPU. I've check and GStreamer is opening
> /dev/dri/card0.
> Also I think I have some sort of CPU frequency changing going on. The
> BogoMIPS in /proc/cpuinfo change depending on CPU loading so that may be
> affecting it. However, I don't seem to have any of the cpufreq modules
> loaded....
you can use "cpufreq-info", and paste the output (apt-get install
cpufrequtils). it gives me this:
$ cpufreq-info
cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to cpuf...@vger.kernel.org, please.
analyzing CPU 0:
driver: omap2plus
CPUs which run at the same hardware frequency: 0 1
CPUs which need to have their frequency coordinated by software: 0 1
maximum transition latency: 300 us.
hardware limits: 350 MHz - 1.20 GHz
available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
available cpufreq governors: conservative, ondemand, userspace,
powersave, performance
current policy: frequency should be within 350 MHz and 1.20 GHz.
The governor "ondemand" may decide which speed to use
within this range.
current CPU frequency is 350 MHz.
cpufreq stats: 350 MHz:99.76%, 700 MHz:0.02%, 920 MHz:0.02%, 1.20
GHz:0.20% (879)
analyzing CPU 1:
driver: omap2plus
CPUs which run at the same hardware frequency: 0 1
CPUs which need to have their frequency coordinated by software: 0 1
maximum transition latency: 300 us.
hardware limits: 350 MHz - 1.20 GHz
available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
available cpufreq governors: conservative, ondemand, userspace,
powersave, performance
current policy: frequency should be within 350 MHz and 1.20 GHz.
The governor "ondemand" may decide which speed to use
within this range.
current CPU frequency is 350 MHz.
cpufreq stats: 350 MHz:99.76%, 700 MHz:0.02%, 920 MHz:0.02%, 1.20
GHz:0.20% (879)
I am running something quite close to what's in the PPA, you should
get something similar.
> Could you suggest some ways of diagnosing this more? Specifically, how could
> I check if the hardware decoding is actually working?
the *easiest* way is to observe interupts during playback, with "itop"
if you use h/w acceleration for video, you will see MAILBOX interupts.
there is 1 interrupt for each frame decoded. so if you decode at
25fps, you will see 25 IT per second ;-)
My results for cpufreq-info are indeed very similar. Specifically they are (and this is when running a 576p h.264 file):
jim@jim-panda:~$ cpufreq-info
cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to cpuf...@vger.kernel.org, please.
analyzing CPU 0:
driver: omap2plus
CPUs which run at the same hardware frequency: 0 1
CPUs which need to have their frequency coordinated by software: 0 1
maximum transition latency: 300 us.
hardware limits: 350 MHz - 1.20 GHz
available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
available cpufreq governors: conservative, ondemand, userspace,
powersave, performance
current policy: frequency should be within 350 MHz and 1.20 GHz.
The governor "ondemand" may decide which speed to use
within this range.
current CPU frequency is 350 MHz.
cpufreq stats: 350 MHz:95.58%, 700 MHz:2.68%, 920 MHz:0.44%, 1.20
GHz:1.30% (850)
analyzing CPU 1:
driver: omap2plus
CPUs which run at the same hardware frequency: 0 1
CPUs which need to have their frequency coordinated by software: 0 1
maximum transition latency: 300 us.
hardware limits: 350 MHz - 1.20 GHz
available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
available cpufreq governors: conservative, ondemand, userspace,
powersave, performance
current policy: frequency should be within 350 MHz and 1.20 GHz.
The governor "ondemand" may decide which speed to use
within this range.
current CPU frequency is 350 MHz.
cpufreq stats: 350 MHz:95.58%, 700 MHz:2.68%, 920 MHz:0.44%, 1.20
GHz:1.30% (850)
I'm getting the 25 interrupts per second on the mailbox:
and generally it's looking pretty much as it should. However, I am still getting the 75% CPU utilisation.
But! Not that the cpufreq-info result above said it was using 350MHz! Maybe that's the give away? Let's try a quick sudo cpufreq-set -g performance and see what happens! Humm... Down to 35%. Now let's try a 720p film: and I'm getting around 100%. That's curious.... Not the 20% we were expecting.
I wonder where all the time is going? Audio decode?
Now that's interesting: switching to a different film in an MKV with H.264 video and DTS audio I'm down to 20% CPU! Hurrah! But with H.264 and AAC audio in a .MP4 I get the high load.
Looks like this is worth noting.
Many thanks for your help in sorting this one,
Jim.
p.s. Now just the black login screen and the losing of debugging on the serial output to solve.....
> On Fri, Sep 7, 2012 at 5:31 PM, Jim Darby<uberscuba...@gmail.com> wrote:
>> OK, it was a H.264 encoded 720p film so it should have been better. I've
>> just checked and it seems pretty consistent. Also with a 576p input I seem
>> to be getting around 75% CPU. I've check and GStreamer is opening
>> /dev/dri/card0.
>> Also I think I have some sort of CPU frequency changing going on. The
>> BogoMIPS in /proc/cpuinfo change depending on CPU loading so that may be
>> affecting it. However, I don't seem to have any of the cpufreq modules
>> loaded....
> you can use "cpufreq-info", and paste the output (apt-get install
> cpufrequtils). it gives me this:
> $ cpufreq-info
> cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009
> Report errors and bugs to cpuf...@vger.kernel.org, please.
> analyzing CPU 0:
> driver: omap2plus
> CPUs which run at the same hardware frequency: 0 1
> CPUs which need to have their frequency coordinated by software: 0 1
> maximum transition latency: 300 us.
> hardware limits: 350 MHz - 1.20 GHz
> available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
> available cpufreq governors: conservative, ondemand, userspace,
> powersave, performance
> current policy: frequency should be within 350 MHz and 1.20 GHz.
> The governor "ondemand" may decide which speed to use
> within this range.
> current CPU frequency is 350 MHz.
> cpufreq stats: 350 MHz:99.76%, 700 MHz:0.02%, 920 MHz:0.02%, 1.20
> GHz:0.20% (879)
> analyzing CPU 1:
> driver: omap2plus
> CPUs which run at the same hardware frequency: 0 1
> CPUs which need to have their frequency coordinated by software: 0 1
> maximum transition latency: 300 us.
> hardware limits: 350 MHz - 1.20 GHz
> available frequency steps: 350 MHz, 700 MHz, 920 MHz, 1.20 GHz
> available cpufreq governors: conservative, ondemand, userspace,
> powersave, performance
> current policy: frequency should be within 350 MHz and 1.20 GHz.
> The governor "ondemand" may decide which speed to use
> within this range.
> current CPU frequency is 350 MHz.
> cpufreq stats: 350 MHz:99.76%, 700 MHz:0.02%, 920 MHz:0.02%, 1.20
> GHz:0.20% (879)
> I am running something quite close to what's in the PPA, you should
> get something similar.
>> Could you suggest some ways of diagnosing this more? Specifically, how could
>> I check if the hardware decoding is actually working?
> the *easiest* way is to observe interupts during playback, with "itop"
> if you use h/w acceleration for video, you will see MAILBOX interupts.
> there is 1 interrupt for each frame decoded. so if you decode at
> 25fps, you will see 25 IT per second ;-)
