[Xenomai] Xenomai-3 booting problems (2)

[Xenomai] Xenomai-3 booting problems (2)

[Leopold Palomo-Avellaneda]

Hi again,

after made the modifications proposed by Jan Kiszka, now it doesn't built and 
the message is:

nvidiafb: unable to setup MTRR

I could try to test in another hardware, but tomorrow.

Regards,

Leopold

-- 
--
Linux User 152692     GPG: 05F4A7A949A2D9AA
Catalonia
-------------------------------------
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150122/12a67aa0/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Leopold Palomo-Avellaneda]

El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold Palomo-Avellaneda va 
escriure:
> Hi again,
> 
> after made the modifications proposed by Jan Kiszka, now it doesn't built
> and the message is:
> 
> nvidiafb: unable to setup MTRR
> 

Hi,

removed nvidiafb.ko module (and all the nvidia graphics stuff) and the box 
booted. It seems that nvidiafb (the non-free nvidia driver) doesn't like 
xenomai-3 or vice-versa.

Are you aware of this?

Leopold 


-- 
--
Linux User 152692     GPG: 05F4A7A949A2D9AA
Catalonia
-------------------------------------
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/7986ac2b/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda wrote:
> El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold Palomo-Avellaneda va 
> escriure:
> > Hi again,
> > 
> > after made the modifications proposed by Jan Kiszka, now it doesn't built
> > and the message is:
> > 
> > nvidiafb: unable to setup MTRR
> > 
> 
> Hi,
> 
> removed nvidiafb.ko module (and all the nvidia graphics stuff) and the box 
> booted. It seems that nvidiafb (the non-free nvidia driver) doesn't like 
> xenomai-3 or vice-versa.
> 
> Are you aware of this?

Were you trying to use the nvidia module compiled for the vanilla
kernel, or did you recompile the nvidia module for the Xenomai
kernel? The former will not work.

Other than that, the nvidia module has been reported to work with
Xenomai, if the I-pipe patch you use contains this patch:
https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2e55bb6014566c62fa083ed9c7b

-- 
					    Gilles.

Re: [Xenomai] Xenomai-3 booting problems (2)

[Leopold Palomo-Avellaneda]

El Divendres, 23 de gener de 2015, a les 08:44:03, Gilles Chanteperdrix va 
escriure:
> On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda wrote:
> > El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold Palomo-Avellaneda
> > va> 
> > escriure:
> > > Hi again,
> > > 
> > > after made the modifications proposed by Jan Kiszka, now it doesn't
> > > built
> > > and the message is:
> > > 
> > > nvidiafb: unable to setup MTRR
> > 
> > Hi,
> > 
> > removed nvidiafb.ko module (and all the nvidia graphics stuff) and the box
> > booted. It seems that nvidiafb (the non-free nvidia driver) doesn't like
> > xenomai-3 or vice-versa.
> > 
> > Are you aware of this?
> 
> Were you trying to use the nvidia module compiled for the vanilla
> kernel, or did you recompile the nvidia module for the Xenomai
> kernel? The former will not work.

the second one. Using debian packages, when you install a new one, it 
recompiles the modules needed (if they are packaged dkms) So, when I installed 
the debian package, all the modules (that had sources) were recompiled.

> Other than that, the nvidia module has been reported to work with
> Xenomai, if the I-pipe patch you use contains this patch:
> https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2e55b
> b6014566c62fa083ed9c7b


The module that failed was the nvidiafb. Not the others. For instance, now I'm 
running 3.16 with xenomai 3.0.0 in my graphical desktop with nvidia driver. 
It's the nvidia framebuffer that fails. 

In debian systems, adding 
blacklist nvidiafb

to/etc/modprobe.d/nvidia-blacklists-nouveau.conf

solve the issue. 

xenomai has some kind of issue tracker?

Leopold

-- 
--
Linux User 152692     GPG: 05F4A7A949A2D9AA
Catalonia
-------------------------------------
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/741c7c4d/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 10:44:24AM +0100, Leopold Palomo-Avellaneda wrote:
> El Divendres, 23 de gener de 2015, a les 08:44:03, Gilles Chanteperdrix va 
> escriure:
> > On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda wrote:
> > > El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold Palomo-Avellaneda
> > > va> 
> > > escriure:
> > > > Hi again,
> > > > 
> > > > after made the modifications proposed by Jan Kiszka, now it doesn't
> > > > built
> > > > and the message is:
> > > > 
> > > > nvidiafb: unable to setup MTRR
> > > 
> > > Hi,
> > > 
> > > removed nvidiafb.ko module (and all the nvidia graphics stuff) and the box
> > > booted. It seems that nvidiafb (the non-free nvidia driver) doesn't like
> > > xenomai-3 or vice-versa.
> > > 
> > > Are you aware of this?
> > 
> > Were you trying to use the nvidia module compiled for the vanilla
> > kernel, or did you recompile the nvidia module for the Xenomai
> > kernel? The former will not work.
> 
> the second one. Using debian packages, when you install a new one, it 
> recompiles the modules needed (if they are packaged dkms) So, when I installed 
> the debian package, all the modules (that had sources) were recompiled.
> 
> > Other than that, the nvidia module has been reported to work with
> > Xenomai, if the I-pipe patch you use contains this patch:
> > https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2e55b
> > b6014566c62fa083ed9c7b
> 
> 
> The module that failed was the nvidiafb. Not the others. For instance, now I'm 
> running 3.16 with xenomai 3.0.0 in my graphical desktop with nvidia driver. 
> It's the nvidia framebuffer that fails. 

AFAIK, nvidiafb is part of the Linux kernel, not of the nvidia
driver. According to this page:
http://wiki.gentoo.org/wiki/NVidia/nvidia-drivers

- the MTRR error is due to the fact that you disabled MTRR in the
kernel configuration, probably because you disabled IOMMU
- using the nvidiafb module with the nvidia proprietary module is
not recommended, as they are known to conflict at times.

> 
> In debian systems, adding 
> blacklist nvidiafb
> 
> to/etc/modprobe.d/nvidia-blacklists-nouveau.conf
> 
> solve the issue. 
> 
> xenomai has some kind of issue tracker?

You are using it.


-- 
					    Gilles.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 811 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/8f333ebb/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 10:44:24AM +0100, Leopold Palomo-Avellaneda wrote:
> El Divendres, 23 de gener de 2015, a les 08:44:03, Gilles Chanteperdrix va 
> escriure:
> > On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda wrote:
> > > El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold Palomo-Avellaneda
> > > va> 
> > > escriure:
> > > > Hi again,
> > > > 
> > > > after made the modifications proposed by Jan Kiszka, now it doesn't
> > > > built
> > > > and the message is:
> > > > 
> > > > nvidiafb: unable to setup MTRR
> > > 
> > > Hi,
> > > 
> > > removed nvidiafb.ko module (and all the nvidia graphics stuff) and the box
> > > booted. It seems that nvidiafb (the non-free nvidia driver) doesn't like
> > > xenomai-3 or vice-versa.
> > > 
> > > Are you aware of this?
> > 
> > Were you trying to use the nvidia module compiled for the vanilla
> > kernel, or did you recompile the nvidia module for the Xenomai
> > kernel? The former will not work.
> 
> the second one. Using debian packages, when you install a new one, it 
> recompiles the modules needed (if they are packaged dkms) So, when I installed 
> the debian package, all the modules (that had sources) were recompiled.
> 
> > Other than that, the nvidia module has been reported to work with
> > Xenomai, if the I-pipe patch you use contains this patch:
> > https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2e55b
> > b6014566c62fa083ed9c7b
> 
> 
> The module that failed was the nvidiafb. Not the others. For instance, now I'm 
> running 3.16 with xenomai 3.0.0 in my graphical desktop with nvidia driver. 
> It's the nvidia framebuffer that fails. 
> 
> In debian systems, adding 
> blacklist nvidiafb
> 
> to/etc/modprobe.d/nvidia-blacklists-nouveau.conf
> 
> solve the issue. 
> 
> xenomai has some kind of issue tracker?

Also, speaking about tracking issues. You did not answer my
question: all these configurations issues you have have nothing
specific to xenomai 3. How come you did not reuse the same
configuration that you used for Xenomai 2.6.4 ?

-- 
					    Gilles.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 811 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/674d5f12/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Leopold Palomo-Avellaneda]

El Divendres, 23 de gener de 2015, a les 11:18:59, Gilles Chanteperdrix va 
escriure:
> On Fri, Jan 23, 2015 at 10:44:24AM +0100, Leopold Palomo-Avellaneda wrote:
> > El Divendres, 23 de gener de 2015, a les 08:44:03, Gilles Chanteperdrix va
> > 
> > escriure:
> > > On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda 
wrote:
> > > > El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold
> > > > Palomo-Avellaneda
> > > > va>
> > > > 
> > > > escriure:
> > > > > Hi again,
> > > > > 
> > > > > after made the modifications proposed by Jan Kiszka, now it doesn't
> > > > > built
> > > > > and the message is:
> > > > > 
> > > > > nvidiafb: unable to setup MTRR
> > > > 
> > > > Hi,
> > > > 
> > > > removed nvidiafb.ko module (and all the nvidia graphics stuff) and the
> > > > box
> > > > booted. It seems that nvidiafb (the non-free nvidia driver) doesn't
> > > > like
> > > > xenomai-3 or vice-versa.
> > > > 
> > > > Are you aware of this?
> > > 
> > > Were you trying to use the nvidia module compiled for the vanilla
> > > kernel, or did you recompile the nvidia module for the Xenomai
> > > kernel? The former will not work.
> > 
> > the second one. Using debian packages, when you install a new one, it
> > recompiles the modules needed (if they are packaged dkms) So, when I
> > installed the debian package, all the modules (that had sources) were
> > recompiled.> 
> > > Other than that, the nvidia module has been reported to work with
> > > Xenomai, if the I-pipe patch you use contains this patch:
> > > https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2
> > > e55b b6014566c62fa083ed9c7b
> > 
> > The module that failed was the nvidiafb. Not the others. For instance, now
> > I'm running 3.16 with xenomai 3.0.0 in my graphical desktop with nvidia
> > driver. It's the nvidia framebuffer that fails.
> 
> AFAIK, nvidiafb is part of the Linux kernel, not of the nvidia
> driver. According to this page:
> http://wiki.gentoo.org/wiki/NVidia/nvidia-drivers

Not exactly. The nvidia driver has two parts: one for the kernel and another 
for the X server. The kernel part has public source (no open source?), and 
it's recompiled as module. The other is closed, binary. 

AFAIK, the frame buffer is to have console in booting time, or console in 
general. It's used when you switch from graphic mode to console mode. But

> - the MTRR error is due to the fact that you disabled MTRR in the
> kernel configuration, probably because you disabled IOMMU

No, I have it activated.

$ grep CONFIG_MTRR /boot/config-3.16.0-xenomai-3.0.0 
CONFIG_MTRR=y
CONFIG_MTRR_SANITIZER=y
CONFIG_MTRR_SANITIZER_ENABLE_DEFAULT=0
CONFIG_MTRR_SANITIZER_SPARE_REG_NR_DEFAULT=1


> - using the nvidiafb module with the nvidia proprietary module is
> not recommended, as they are known to conflict at times.

It could have conflicts, it's true. Because the two modules try to claim the 
same device. I should have the test installing only one and see what happens.

> > In debian systems, adding
> > blacklist nvidiafb
> > 
> > to/etc/modprobe.d/nvidia-blacklists-nouveau.conf
> > 
> > solve the issue.
> > 
> > xenomai has some kind of issue tracker?
> 
> You are using it.

I love this kind of issue tracker :-)


Leopold

-- 
--
Linux User 152692     GPG: 05F4A7A949A2D9AA
Catalonia
-------------------------------------
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/b2503f38/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Leopold Palomo-Avellaneda]

El Divendres, 23 de gener de 2015, a les 11:24:26, Gilles Chanteperdrix va 
escriure:

> > xenomai has some kind of issue tracker?
> 
> Also, speaking about tracking issues. You did not answer my
> question: all these configurations issues you have have nothing
> specific to xenomai 3. How come you did not reuse the same
> configuration that you used for Xenomai 2.6.4 ?

No Gilles. I think that I wrote it before (as I said ;-)

I copied the same configuration from xenomai-2.6.4 to xenomai-3.0.0. Xenomai 3 
introduced some differences. I have attached the diffs.

The fb issue has appeared because I test the kernels in a box without graphic 
environtment but today I tested in my box (with graphic environtment). I 
prefer the nvidia cards in general for the labs because I have less problems 
than the ati ones.

Debian drops the nvidiafb, so is the reason why I have never seen this 
conflict. That's the reason,

Leopold



-- 
--
Linux User 152692     GPG: 05F4A7A949A2D9AA
Catalonia
-------------------------------------
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
-------------- next part --------------
3c3
< # Linux/x86_64 3.16.0-xenomai-2.6.4.git.ae4b2edc Kernel Configuration
---
> # Linux/x86_64 3.16.0-xenomai-3.0.0 Kernel Configuration
37d36
< CONFIG_HAVE_INTEL_TXT=y
357a357,359
> CONFIG_XENOMAI=y
> CONFIG_XENO_ARCH_FPU=y
> CONFIG_XENO_ARCH_SYS3264=y
360c362
< # Real-time sub-system
---
> # Core features
362,368d363
< CONFIG_XENOMAI=y
< CONFIG_XENO_GENERIC_STACKPOOL=y
< CONFIG_XENO_FASTSYNCH=y
< CONFIG_XENO_OPT_NUCLEUS=y
< CONFIG_XENO_OPT_PERVASIVE=y
< # CONFIG_XENO_OPT_PRIOCPL is not set
< CONFIG_XENO_OPT_PIPELINE_HEAD=y
370,378d364
< CONFIG_XENO_OPT_PIPE=y
< CONFIG_XENO_OPT_MAP=y
< CONFIG_XENO_OPT_VFILE=y
< CONFIG_XENO_OPT_PIPE_NRDEV=512
< CONFIG_XENO_OPT_REGISTRY_NRSLOTS=512
< CONFIG_XENO_OPT_SYS_HEAPSZ=512
< CONFIG_XENO_OPT_SYS_STACKPOOLSZ=128
< CONFIG_XENO_OPT_SEM_HEAPSZ=12
< CONFIG_XENO_OPT_GLOBAL_SEM_HEAPSZ=12
380,388d365
< CONFIG_XENO_OPT_DEBUG=y
< CONFIG_XENO_OPT_DEBUG_NUCLEUS=y
< CONFIG_XENO_OPT_DEBUG_XNLOCK=y
< # CONFIG_XENO_OPT_DEBUG_QUEUES is not set
< # CONFIG_XENO_OPT_DEBUG_REGISTRY is not set
< # CONFIG_XENO_OPT_DEBUG_TIMERS is not set
< CONFIG_XENO_OPT_DEBUG_SYNCH_RELAX=y
< CONFIG_XENO_OPT_WATCHDOG=y
< CONFIG_XENO_OPT_WATCHDOG_TIMEOUT=4
390,397c367,368
< CONFIG_XENO_OPT_SELECT=y
< CONFIG_XENO_OPT_HOSTRT=y
< 
< #
< # Timing
< #
< CONFIG_XENO_OPT_TIMING_PERIODIC=y
< CONFIG_XENO_OPT_TIMING_VIRTICK=1000
---
> CONFIG_XENO_OPT_RR_QUANTUM=1000
> CONFIG_XENO_OPT_AUTOTUNE=m
399,402d369
< 
< #
< # Scalability
< #
406c373,376
< # CONFIG_XENO_OPT_TIMER_WHEEL is not set
---
> CONFIG_XENO_OPT_HOSTRT=y
> CONFIG_XENO_OPT_PIPE=y
> CONFIG_XENO_OPT_MAP=y
> CONFIG_XENO_OPT_VFILE=y
409,412c379
< # Machine
< #
< CONFIG_XENO_HW_FPU=y
< 
---
> # Sizes and static limits
414,457c381,394
< # Interfaces
< #
< CONFIG_XENO_SKIN_NATIVE=y
< CONFIG_XENO_OPT_NATIVE_PERIOD=0
< CONFIG_XENO_OPT_NATIVE_PIPE=y
< CONFIG_XENO_OPT_NATIVE_PIPE_BUFSZ=4096
< CONFIG_XENO_OPT_NATIVE_SEM=y
< CONFIG_XENO_OPT_NATIVE_EVENT=y
< CONFIG_XENO_OPT_NATIVE_MUTEX=y
< CONFIG_XENO_OPT_NATIVE_COND=y
< CONFIG_XENO_OPT_NATIVE_QUEUE=y
< CONFIG_XENO_OPT_NATIVE_BUFFER=y
< CONFIG_XENO_OPT_NATIVE_HEAP=y
< CONFIG_XENO_OPT_NATIVE_ALARM=y
< CONFIG_XENO_OPT_NATIVE_MPS=y
< # CONFIG_XENO_OPT_NATIVE_INTR is not set
< CONFIG_XENO_OPT_DEBUG_NATIVE=y
< CONFIG_XENO_SKIN_POSIX=y
< CONFIG_XENO_OPT_POSIX_PERIOD=0
< CONFIG_XENO_OPT_POSIX_REGISTRY_NRSLOTS=64
< CONFIG_XENO_OPT_POSIX_REGISTRY_NRDESCS=128
< CONFIG_XENO_OPT_POSIX_NRTIMERS=128
< # CONFIG_XENO_OPT_POSIX_SHM is not set
< # CONFIG_XENO_OPT_POSIX_INTR is not set
< CONFIG_XENO_OPT_POSIX_SELECT=y
< CONFIG_XENO_OPT_DEBUG_POSIX=y
< CONFIG_XENO_SKIN_PSOS=m
< CONFIG_XENO_OPT_PSOS_PERIOD=1000
< CONFIG_XENO_OPT_DEBUG_PSOS=y
< CONFIG_XENO_SKIN_UITRON=m
< CONFIG_XENO_OPT_UITRON_PERIOD=1000
< CONFIG_XENO_OPT_DEBUG_UITRON=y
< CONFIG_XENO_SKIN_VRTX=m
< CONFIG_XENO_OPT_VRTX_PERIOD=1000
< CONFIG_XENO_SKIN_VXWORKS=m
< CONFIG_XENO_OPT_VXWORKS_PERIOD=1000
< CONFIG_XENO_OPT_DEBUG_VXWORKS=y
< # CONFIG_XENO_OPT_NOWARN_DEPRECATED is not set
< CONFIG_XENO_SKIN_RTDM=y
< CONFIG_XENO_OPT_RTDM_PERIOD=0
< CONFIG_XENO_OPT_RTDM_FILDES=512
< CONFIG_XENO_OPT_RTDM_SELECT=y
< # CONFIG_XENO_OPT_DEBUG_RTDM is not set
< CONFIG_XENO_OPT_DEBUG_RTDM_APPL=y
---
> CONFIG_XENO_OPT_PIPE_NRDEV=512
> CONFIG_XENO_OPT_REGISTRY_NRSLOTS=512
> CONFIG_XENO_OPT_SYS_HEAPSZ=512
> CONFIG_XENO_OPT_PRIVATE_HEAPSZ=32
> CONFIG_XENO_OPT_SHARED_HEAPSZ=32
> CONFIG_XENO_OPT_NRTIMERS=128
> CONFIG_XENO_OPT_DEBUG=y
> # CONFIG_XENO_OPT_DEBUG_COBALT is not set
> # CONFIG_XENO_OPT_DEBUG_CONTEXT is not set
> CONFIG_XENO_OPT_DEBUG_LOCKING=y
> # CONFIG_XENO_OPT_DEBUG_USER is not set
> # CONFIG_XENO_OPT_DEBUG_TRACE_RELAX is not set
> CONFIG_XENO_OPT_WATCHDOG=y
> CONFIG_XENO_OPT_WATCHDOG_TIMEOUT=4
461a399,400
> CONFIG_XENO_OPT_RTDM_COMPAT_DEVNODE=y
> CONFIG_XENO_DRIVERS_AUTOTUNE=m
477,478d415
< CONFIG_XENO_DRIVERS_KLATENCY=m
< CONFIG_XENO_DRIVERS_IRQBENCH=m
494d430
< CONFIG_XENO_DRIVERS_CAN_FLEXCAN=m
506a443,512
> # RTnet
> #
> CONFIG_XENO_DRIVERS_NET=m
> 
> #
> # Protocol Stack
> #
> 
> #
> # Stack parameters
> #
> CONFIG_XENO_DRIVERS_NET_RX_FIFO_SIZE=32
> # CONFIG_XENO_DRIVERS_NET_ETH_P_ALL is not set
> CONFIG_XENO_DRIVERS_NET_RTWLAN=y
> 
> #
> # Protocols
> #
> CONFIG_XENO_DRIVERS_NET_RTIPV4=m
> CONFIG_XENO_DRIVERS_NET_RTIPV4_ICMP=y
> CONFIG_XENO_DRIVERS_NET_RTIPV4_HOST_ROUTES=32
> # CONFIG_XENO_DRIVERS_NET_RTIPV4_NETROUTING is not set
> # CONFIG_XENO_DRIVERS_NET_RTIPV4_ROUTER is not set
> CONFIG_XENO_DRIVERS_NET_RTIPV4_UDP=m
> CONFIG_XENO_DRIVERS_NET_RTPACKET=m
> CONFIG_XENO_DRIVERS_NET_RTMAC=m
> CONFIG_XENO_DRIVERS_NET_TDMA=m
> CONFIG_XENO_DRIVERS_NET_TDMA_MASTER=y
> CONFIG_XENO_DRIVERS_NET_NOMAC=m
> CONFIG_XENO_DRIVERS_NET_RTCFG=m
> # CONFIG_XENO_DRIVERS_NET_RTCFG_DEBUG is not set
> 
> #
> # Drivers
> #
> 
> #
> # Common PCI Drivers
> #
> CONFIG_XENO_DRIVERS_NET_DRV_PCNET32=m
> CONFIG_XENO_DRIVERS_NET_DRV_TULIP=m
> CONFIG_XENO_DRIVERS_NET_DRV_EEPRO100=m
> CONFIG_XENO_DRIVERS_NET_DRV_EEPRO100_CMDTIMEOUT=20
> # CONFIG_XENO_DRIVERS_NET_DRV_EEPRO100_DBG is not set
> CONFIG_XENO_DRIVERS_NET_DRV_E1000=m
> CONFIG_XENO_DRIVERS_NET_DRV_E1000E=m
> CONFIG_XENO_DRIVERS_NET_DRV_NATSEMI=m
> CONFIG_XENO_DRIVERS_NET_DRV_8139=m
> CONFIG_XENO_DRIVERS_NET_DRV_VIA_RHINE=m
> CONFIG_XENO_DRIVERS_NET_DRV_IGB=m
> CONFIG_XENO_DRIVERS_NET_DRV_R8169=m
> 
> #
> # Misc Drivers
> #
> CONFIG_XENO_DRIVERS_NET_DRV_LOOPBACK=m
> CONFIG_XENO_DRIVERS_NET_DRV_SMC91111=m
> CONFIG_XENO_DRIVERS_NET_EXP_DRIVERS=y
> CONFIG_XENO_DRIVERS_NET_DRV_3C59X=m
> CONFIG_XENO_DRIVERS_NET_DRV_E1000_NEW=m
> CONFIG_XENO_DRIVERS_NET_DRV_RT2500=m
> 
> #
> # Add-Ons
> #
> CONFIG_XENO_DRIVERS_NET_ADDON_RTCAP=m
> CONFIG_XENO_DRIVERS_NET_ADDON_PROXY=m
> # CONFIG_XENO_DRIVERS_NET_ADDON_PROXY_ARP is not set
> 
> #
510a517
> # CONFIG_XENO_DRIVERS_ANALOGY_DEBUG_FTRACE is not set
530a538,547
> 
> #
> # UDD support
> #
> CONFIG_XENO_DRIVERS_UDD=m
> CONFIG_XENO_VERSION_MAJOR=2
> CONFIG_XENO_VERSION_MINOR=99
> CONFIG_XENO_REVISION_LEVEL=10
> CONFIG_XENO_VERSION_STRING="3.0-rc2"
> CONFIG_XENO_VERSION_NAME="Exact Zero"
538d554
< CONFIG_X86_X2APIC=y
579c595
< CONFIG_IPIPE_LEGACY=y
---
> # CONFIG_IPIPE_LEGACY is not set
582d597
< CONFIG_IPIPE_WANT_APIREV_2=y
585d599
< CONFIG_IPIPE_DELAYED_ATOMICSW=y
5221d5234
< # CONFIG_INFINIBAND_USNIC is not set
5367,5370d5379
< CONFIG_VFIO_IOMMU_TYPE1=m
< CONFIG_VFIO=m
< CONFIG_VFIO_PCI=m
< CONFIG_VFIO_PCI_VGA=y
5764,5773c5773
< CONFIG_IOMMU_API=y
< CONFIG_IOMMU_SUPPORT=y
< CONFIG_AMD_IOMMU=y
< # CONFIG_AMD_IOMMU_STATS is not set
< CONFIG_AMD_IOMMU_V2=y
< CONFIG_DMAR_TABLE=y
< CONFIG_INTEL_IOMMU=y
< # CONFIG_INTEL_IOMMU_DEFAULT_ON is not set
< CONFIG_INTEL_IOMMU_FLOPPY_WA=y
< CONFIG_IRQ_REMAP=y
---
> # CONFIG_IOMMU_SUPPORT is not set
6530d6529
< # CONFIG_INTEL_TXT is not set
6726d6724
< CONFIG_KVM_DEVICE_ASSIGNMENT=y
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/6dcb6c92/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 12:30:02PM +0100, Leopold Palomo-Avellaneda wrote:
> El Divendres, 23 de gener de 2015, a les 11:18:59, Gilles Chanteperdrix va 
> escriure:
> > On Fri, Jan 23, 2015 at 10:44:24AM +0100, Leopold Palomo-Avellaneda wrote:
> > > El Divendres, 23 de gener de 2015, a les 08:44:03, Gilles Chanteperdrix va
> > > 
> > > escriure:
> > > > On Fri, Jan 23, 2015 at 08:36:14AM +0100, Leopold Palomo-Avellaneda 
> wrote:
> > > > > El Dijous, 22 de gener de 2015, a les 16:09:19, Leopold
> > > > > Palomo-Avellaneda
> > > > > va>
> > > > > 
> > > > > escriure:
> > > > > > Hi again,
> > > > > > 
> > > > > > after made the modifications proposed by Jan Kiszka, now it doesn't
> > > > > > built
> > > > > > and the message is:
> > > > > > 
> > > > > > nvidiafb: unable to setup MTRR
> > > > > 
> > > > > Hi,
> > > > > 
> > > > > removed nvidiafb.ko module (and all the nvidia graphics stuff) and the
> > > > > box
> > > > > booted. It seems that nvidiafb (the non-free nvidia driver) doesn't
> > > > > like
> > > > > xenomai-3 or vice-versa.
> > > > > 
> > > > > Are you aware of this?
> > > > 
> > > > Were you trying to use the nvidia module compiled for the vanilla
> > > > kernel, or did you recompile the nvidia module for the Xenomai
> > > > kernel? The former will not work.
> > > 
> > > the second one. Using debian packages, when you install a new one, it
> > > recompiles the modules needed (if they are packaged dkms) So, when I
> > > installed the debian package, all the modules (that had sources) were
> > > recompiled.> 
> > > > Other than that, the nvidia module has been reported to work with
> > > > Xenomai, if the I-pipe patch you use contains this patch:
> > > > https://git.xenomai.org/ipipe.git/commit/?h=ipipe-3.16&id=bc5bea6f134de2
> > > > e55b b6014566c62fa083ed9c7b
> > > 
> > > The module that failed was the nvidiafb. Not the others. For instance, now
> > > I'm running 3.16 with xenomai 3.0.0 in my graphical desktop with nvidia
> > > driver. It's the nvidia framebuffer that fails.
> > 
> > AFAIK, nvidiafb is part of the Linux kernel, not of the nvidia
> > driver. According to this page:
> > http://wiki.gentoo.org/wiki/NVidia/nvidia-drivers
> 
> Not exactly. The nvidia driver has two parts: one for the kernel and another 
> for the X server.


> The kernel part has public source (no open source?), and 
> it's recompiled as module. The other is closed, binary. 

No the kernel part contains some glue with a big pre-compiled
proprietary binary.

> 
> AFAIK, the frame buffer is to have console in booting time, or console in 
> general. It's used when you switch from graphic mode to console mode. But

You are mistaken. The nvidiafb driver is part of the Linux kernel,
see:
http://git.xenomai.org/ipipe.git/tree/drivers/video/fbdev/nvidia/Makefile

In this file, you can see the copyright:
http://git.xenomai.org/ipipe.git/tree/drivers/video/fbdev/nvidia/nvidia.c

Not by Nvidia. Yes nvidiafb is a framebuffer driver, I knew that,
only by the name. But no, it is not part of the (binary) drivers
made by nvidia, so it is a third thing in addition to the nvidia.ko
kernel-space support provided by Nvidia, and the user-space
distribution (which contains a module for the X server, but also an
opengl implementation).

And the gentoo page seems to indicate that the framebuffer driver
may conflict with the nvidia stuff (whether kernel-space or
user-space does not really matter).

> 
> > - the MTRR error is due to the fact that you disabled MTRR in the
> > kernel configuration, probably because you disabled IOMMU
> 
> No, I have it activated.
> 
> $ grep CONFIG_MTRR /boot/config-3.16.0-xenomai-3.0.0 
> CONFIG_MTRR=y
> CONFIG_MTRR_SANITIZER=y
> CONFIG_MTRR_SANITIZER_ENABLE_DEFAULT=0
> CONFIG_MTRR_SANITIZER_SPARE_REG_NR_DEFAULT=1

You have not read the page I sent you. It clearly indicates that on
the amd64 architecture, you have to enable IOMMU to get MTRR working
with the nvidiafb module. I do not know whether this is true, but
this is what the page says.

> 
> 
> > - using the nvidiafb module with the nvidia proprietary module is
> > not recommended, as they are known to conflict at times.
> 
> It could have conflicts, it's true. Because the two modules try to claim the 
> same device. I should have the test installing only one and see what happens.
> 

Obviously.



-- 
					    Gilles.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 811 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/c323b4f7/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 12:43:05PM +0100, Leopold Palomo-Avellaneda wrote:
> El Divendres, 23 de gener de 2015, a les 11:24:26, Gilles Chanteperdrix va 
> escriure:
> 
> > > xenomai has some kind of issue tracker?
> > 
> > Also, speaking about tracking issues. You did not answer my
> > question: all these configurations issues you have have nothing
> > specific to xenomai 3. How come you did not reuse the same
> > configuration that you used for Xenomai 2.6.4 ?
> 
> No Gilles. I think that I wrote it before (as I said ;-)
> 
> I copied the same configuration from xenomai-2.6.4 to xenomai-3.0.0. Xenomai 3 
> introduced some differences. I have attached the diffs.

The CONFIG_XENO* differences and the CONFIG_IPIPE* differences
are obviously due to Xenomai version change. The rest is not.

> 
> The fb issue has appeared because I test the kernels in a box without graphic 
> environtment but today I tested in my box (with graphic environtment). I 
> prefer the nvidia cards in general for the labs because I have less problems 
> than the ati ones.

I prefer the Intel ones: the drivers are free, as in free beer, are
mainlined into all the projects involved (linux, libdrm, mesa,
libva, xorg, etc..) have a working hardware acceleration with good
performances (maybe not for bleeding edge games, but OK for usual
3D needs and for decompressing H264 and displaying video).

> < CONFIG_VFIO_IOMMU_TYPE1=m
> < CONFIG_VFIO=m
> < CONFIG_VFIO_PCI=m
> < CONFIG_VFIO_PCI_VGA=y
> 5764,5773c5773
> < CONFIG_IOMMU_API=y
> < CONFIG_IOMMU_SUPPORT=y
> < CONFIG_AMD_IOMMU=y
> < # CONFIG_AMD_IOMMU_STATS is not set
> < CONFIG_AMD_IOMMU_V2=y
> < CONFIG_DMAR_TABLE=y
> < CONFIG_INTEL_IOMMU=y
> < # CONFIG_INTEL_IOMMU_DEFAULT_ON is not set
> < CONFIG_INTEL_IOMMU_FLOPPY_WA=y
> < CONFIG_IRQ_REMAP=y
> ---
> > # CONFIG_IOMMU_SUPPORT is not set
> 6530d6529
> < # CONFIG_INTEL_TXT is not set
> 6726d6724
> < CONFIG_KVM_DEVICE_ASSIGNMENT=y


-- 
					    Gilles.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 811 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20150123/7231044f/attachment.sig>

Re: [Xenomai] Xenomai-3 booting problems (2)

[Daniele Nicolodi]

On 23/01/15 15:08, Gilles Chanteperdrix wrote:
> On Fri, Jan 23, 2015 at 12:43:05PM +0100, Leopold Palomo-Avellaneda wrote:
>> The fb issue has appeared because I test the kernels in a box without graphic 
>> environtment but today I tested in my box (with graphic environtment). I 
>> prefer the nvidia cards in general for the labs because I have less problems 
>> than the ati ones.
> 
> I prefer the Intel ones: the drivers are free, as in free beer, are

Probably you meant "free as in free speech" here :)

Cheers,
Daniele

Re: [Xenomai] Xenomai-3 booting problems (2)

[Gilles Chanteperdrix]

On Fri, Jan 23, 2015 at 03:16:52PM +0100, Daniele Nicolodi wrote:
> On 23/01/15 15:08, Gilles Chanteperdrix wrote:
> > On Fri, Jan 23, 2015 at 12:43:05PM +0100, Leopold Palomo-Avellaneda wrote:
> >> The fb issue has appeared because I test the kernels in a box without graphic 
> >> environtment but today I tested in my box (with graphic environtment). I 
> >> prefer the nvidia cards in general for the labs because I have less problems 
> >> than the ati ones.
> > 
> > I prefer the Intel ones: the drivers are free, as in free beer, are
> 
> Probably you meant "free as in free speech" here :)

Yes, sorry.

-- 
					    Gilles.