dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Sat Jan 12, 2013 11:19 am

Rolinh wrote: Just to confirm what pepedog said: I tried a testing image of archlinuxarm that pepedog created with the latest firmware and I am unable to boot. Using the old bootcode.bin from archlinuxarm's September image with this image, it works fine. However this problem only happens when using this SD card: SANDISK Ultra 16G. I tried the new image with a SANDISK Extreme 16G and a Sony UHS-1 16GB and it works fine. All of theses SD cards are class 10. I am not sure what is going on here.
Have you enabled CMA?
Any vchiq related messages when booting?

Rolinh
Posts: 3
Joined: Sat Jan 12, 2013 9:49 am

Re: Updated GPU firmware

Sat Jan 12, 2013 12:32 pm

I haven't touched anything from pepedog's image.

Code: Select all

[robin@heimdall ~] % grep -v '^$' /boot/config.txt |grep -v '^ *#'
gpu_mem_512=316
gpu_mem_256=128
cma_lwm=16
cma_hwm=32
cma_offline_start=16
As for the messages at boot, I can't provide any information since it won't even display something on the screen. Also, I forgot to mention that I am using the latest revision of model B.

Many thanks :)

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Sat Jan 12, 2013 4:20 pm

Rolinh wrote:I haven't touched anything from pepedog's image.

Code: Select all

[robin@heimdall ~] % grep -v '^$' /boot/config.txt |grep -v '^ *#'
gpu_mem_512=316
gpu_mem_256=128
cma_lwm=16
cma_hwm=32
cma_offline_start=16
They are not quite the recommended values (although I don't think that is your problem).
http://www.raspberrypi.org/phpBB3/viewt ... 25#p223549

I suspect you don't have this fix:
https://github.com/raspberrypi/linux/co ... 464e9c629e

Can you (or Pepedog) confirm that?

Rolinh
Posts: 3
Joined: Sat Jan 12, 2013 9:49 am

Re: Updated GPU firmware

Sat Jan 12, 2013 5:20 pm

Looking in the archlinuxarm repo (https://github.com/archlinuxarm/PKGBUIL ... aspberrypi), I don't not think the patch has been applied. However, pepedog may have applied it but I just don't have any idea.

pepedog
Posts: 1043
Joined: Fri Oct 07, 2011 9:55 am

Re: Updated GPU firmware

Sat Jan 12, 2013 6:56 pm

Our source is https://github.com/raspberrypi/linux where the patch is applied
Kernel and firmware are just 2 days old, no commits since clone
Incase the question of what kernel config crops up it's here https://github.com/archlinuxarm/PKGBUIL ... ypi/config

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Sat Jan 12, 2013 7:12 pm

Rolinh wrote:I haven't touched anything from pepedog's image.

Code: Select all

[robin@heimdall ~] % grep -v '^$' /boot/config.txt |grep -v '^ *#'
gpu_mem_512=316
gpu_mem_256=128
cma_lwm=16
cma_hwm=32
cma_offline_start=16
Can you remove those lines from config.txt and confirm that the board boots okay?
(CMA is still experimental, and disabled by default in Raspbian, so there may be problems)

pepedog
Posts: 1043
Joined: Fri Oct 07, 2011 9:55 am

Re: Updated GPU firmware

Sat Jan 12, 2013 7:25 pm

He already tried that. Those settings do work on yet-another-card though

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

Re: Updated GPU firmware

Sun Jan 13, 2013 11:26 am

To check if CMA is enabled:

Code: Select all

cat /proc/vc-cma
Videocore CMA:
   Base       : 0a000000
   Length     : 14c00000
   Initial    : 00000000
   Chunk size : 00040000
   Chunks     : 1328 (348127232 bytes)
   Used       :    0 (0 bytes)
   Reserved   :    0 (0 bytes)
I ok?

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

Re: Updated GPU firmware

Mon Jan 14, 2013 9:04 am

I don't know why I cannot edit my post sometimes.
About the raspberry hangs after firmware update this morning I restarted from a fresh install and I have to reboot twice, it's the 3rd time: shutdown -r now and the firs time the raspberry hangs as I posted in the screenshot, I unplug replug the power and it works fine.

It happened again hangs and booted fine after unplug/plug

I'm running XBMC now, these lines in syslog says that CMA is working and it did some memory "moving"?

Code: Select all

Jan 14 09:34:45 raspberrypi kernel: [  500.111180] Mem-info:
Jan 14 09:34:45 raspberrypi kernel: [  500.111189] Normal per-cpu:
Jan 14 09:34:45 raspberrypi kernel: [  500.111199] CPU    0: hi:  186, btch:  31 usd:  34
Jan 14 09:34:45 raspberrypi kernel: [  500.111224] active_anon:5496 inactive_anon:5567 isolated_anon:0
Jan 14 09:34:45 raspberrypi kernel: [  500.111224]  active_file:6595 inactive_file:17313 isolated_file:0
Jan 14 09:34:45 raspberrypi kernel: [  500.111224]  unevictable:0 dirty:954 writeback:0 unstable:0
Jan 14 09:34:45 raspberrypi kernel: [  500.111224]  free:67333 slab_reclaimable:885 slab_unreclaimable:1141
Jan 14 09:34:45 raspberrypi kernel: [  500.111224]  mapped:6482 shmem:144 pagetables:387 bounce:0
Jan 14 09:34:45 raspberrypi kernel: [  500.111276] Normal free:269332kB min:8192kB low:10240kB high:12288kB active_anon:21984kB inactive_anon:22268kB active_file:26380kB inactive_file:69252kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:483488kB mlocked:0kB dirty:3816kB writeback:0kB mapped:25928kB shmem:576kB slab_reclaimable:3540kB slab_unreclaimable:4564kB kernel_stack:1552kB pagetables:1548kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? no
Jan 14 09:34:45 raspberrypi kernel: [  500.111289] lowmem_reserve[]: 0 0
Jan 14 09:34:45 raspberrypi kernel: [  500.111307] Normal: 8309*4kB 8310*8kB 8309*16kB 1146*32kB 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 269332kB
Jan 14 09:34:45 raspberrypi kernel: [  500.111354] 24054 total pagecache pages
Jan 14 09:34:45 raspberrypi kernel: [  500.111364] 0 pages in swap cache
Jan 14 09:34:45 raspberrypi kernel: [  500.111373] Swap cache stats: add 0, delete 0, find 0/0
Jan 14 09:34:45 raspberrypi kernel: [  500.111381] Free swap  = 246780kB
Jan 14 09:34:45 raspberrypi kernel: [  500.111387] Total swap = 246780kB
Jan 14 09:34:45 raspberrypi kernel: [  500.138169] 121856 pages of RAM
Jan 14 09:34:45 raspberrypi kernel: [  500.138179] 67720 free pages
Jan 14 09:34:45 raspberrypi kernel: [  500.138185] 2716 reserved pages
Jan 14 09:34:45 raspberrypi kernel: [  500.138191] 2026 slab pages
Jan 14 09:34:45 raspberrypi kernel: [  500.138197] 27155 pages shared
Jan 14 09:34:45 raspberrypi kernel: [  500.138204] 0 pages swap cached
Jan 14 09:34:45 raspberrypi kernel: [  500.138237] smsc95xx 1-1.1:1.0: eth0: kevent 2 may have been dropped
Jan 14 09:34:45 raspberrypi kernel: [  500.138519] smsc95xx 1-1.1:1.0: eth0: kevent 2 may have been dropped

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Mon Jan 14, 2013 11:18 am

mcgyver83 wrote:I don't know why I cannot edit my post sometimes.
About the raspberry hangs after firmware update this morning I restarted from a fresh install and I have to reboot twice, it's the 3rd time: shutdown -r now and the firs time the raspberry hangs as I posted in the screenshot, I unplug replug the power and it works fine.
Looking at the code the last line of your screenshot shouldn't be possible with latest firmware without printing:
vchiq_initialise: videocore not initialized
or
vchiq_initialise: videocore initialized after %d retries
So, can you report output of

Code: Select all

uname -a
vcgencmd version
to confirm last update worked. Can you also show output of:

Code: Select all

vcgencmd get_config int
vcgencmd get_config str
cat /proc/cmdline
Also adding boot_delay=1 (or higher number) to config.txt make be worth a try. If you are overclocking, then try disabling it.

User avatar
RaTTuS
Posts: 10498
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
Contact: Twitter YouTube

Re: Updated GPU firmware

Mon Jan 14, 2013 11:23 am

mcgyver83 wrote:I don't know why I cannot edit my post sometimes.
you get 1/2 an hour after posting to make an edit - after that a new post is needed
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

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

Re: Updated GPU firmware

Mon Jan 14, 2013 11:47 am

dom wrote:
mcgyver83 wrote:I don't know why I cannot edit my post sometimes.
About the raspberry hangs after firmware update this morning I restarted from a fresh install and I have to reboot twice, it's the 3rd time: shutdown -r now and the firs time the raspberry hangs as I posted in the screenshot, I unplug replug the power and it works fine.
Looking at the code the last line of your screenshot shouldn't be possible with latest firmware without printing:
vchiq_initialise: videocore not initialized
or
vchiq_initialise: videocore initialized after %d retries
So, can you report output of

Code: Select all

uname -a
vcgencmd version
to confirm last update worked. Can you also show output of:

Code: Select all

vcgencmd get_config int
vcgencmd get_config str
cat /proc/cmdline
Also adding boot_delay=1 (or higher number) to config.txt make be worth a try. If you are overclocking, then try disabling it.

Code: Select all

pi@raspberrypi ~ $ uname -a
Linux raspberrypi 3.6.11+ #352 PREEMPT Wed Jan 9 17:16:53 GMT 2013 armv6l GNU/Linux
pi@raspberrypi ~ $ vcgencmd version
Jan  9 2013 17:27:35 
Copyright (c) 2012 Broadcom
version 361556 (release)

Code: Select all

pi@raspberrypi ~ $ vcgencmd get_config int
temp_limit=85
cma_lwm=16
cma_hwm=32
cma_offline_start=16
pi@raspberrypi ~ $ vcgencmd get_config str
pi@raspberrypi ~ $ cat /proc/cmdline
dma.dmachans=0x7f35 bcm2708_fb.fbwidth=1824 bcm2708_fb.fbheight=984 bcm2708.boardrev=0xf bcm2708.serial=0x5db1dbab smsc95xx.macaddr=B8:27:EB:B1:DB:AB sdhci-bcm2708.emmc_clock_freq=100000000 vc-cma-mem=0/0x14c00000@0xa000000 mem=0x9000000@0x0 mem=0x14c00000@0xa000000 vc_mem.mem_base=0x1ec00000 vc_mem.mem_size=0x20000000  dwc_otg.lpm_enable=0 console=ttyAMA0,115200 kgdboc=ttyAMA0,115200 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait coherent_pool=6M  smsc95xx.turbo_mode=N
pi@raspberrypi ~ $ 
No overclock now :D
I'm going to try with "boot_delay=1", it's only annoying because if I unplug/replug the power a couple of times it boots.let's try.

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

Re: Updated GPU firmware

Mon Jan 14, 2013 4:44 pm

I did again another rpi-update, to be sure that I have all updated.
I had a couple of time the same issue but after it disappeared.
Encouraged I tried with overclock: if I enable overclock (very confident value, small overclock) the previous hang happen again; if I remove the oc it boots fine....

I don't know why but it's fine now...in the week I'm going to try again overclock.

pepedog
Posts: 1043
Joined: Fri Oct 07, 2011 9:55 am

Re: Updated GPU firmware

Thu Jan 17, 2013 3:29 pm

I have several people using arch complaing that on update to latest kernel and firmware that they can't boot
They all have SanDisk SDHC Ultra Class 10 of various sizes, think they are all UHS (only one confirmed so far) with this marking Image
http://en.wikipedia.org/wiki/Secure_Dig ... peed_Class
Aparently getting an older bootcode.bin helps to boot

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Thu Jan 17, 2013 5:56 pm

pepedog wrote:Aparently getting an older bootcode.bin helps to boot
Can you identify the newest bootcode.bin that still works (and so the first bootcode.bin with the problem) ?
With that I could look through the changes and perhaps diagnose the problem.

pepedog
Posts: 1043
Joined: Fri Oct 07, 2011 9:55 am

Re: Updated GPU firmware

Thu Jan 17, 2013 6:35 pm

I'm now getting that latest firmware (via arch update) that the problem has gone now, with version 362704
Problems occurred about 20 Dec and I think that upgrade had firmware 2 weeks old, they we resorting to September blobs with results like not full ram available, but all this is in the past and there are happy faces at last
Thanks Dom

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: Updated GPU firmware

Thu Jan 17, 2013 10:49 pm

No better results than the previous firmware for me. It never worked following a re-boot and maybe 1 of 8 (if that) tries from a power up. If there are new suggested settings for cmdline.txt/config.txt, I'll be happy to try them out. Doesn't seem to matter if I OC to 1100 or leave it at 700. FW 362704 KERNEL3.6.11+ #358
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

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

Re: Updated GPU firmware

Fri Jan 18, 2013 8:10 am

I had:

Code: Select all

Linux raspberrypi 3.6.11+ #354 PREEMPT Sun Jan 13 16:13:26 GMT 2013 armv6l GNU/Linux

Code: Select all

Jan 13 2013 16:24:29 
Copyright (c) 2012 Broadcom
version 362371 (release)
now after and rpi-update

Code: Select all

Linux raspberrypi 3.6.11+ #358 PREEMPT Tue Jan 15 00:45:33 GMT 2013 armv6l GNU/Linux

Code: Select all

Jan 15 2013 12:54:01 
Copyright (c) 2012 Broadcom
version 362704 (release)
I rebooted after the update and it runs fine.

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

Re: Updated GPU firmware

Fri Jan 18, 2013 10:27 am

I also retried after this update but overclock still block the boot at the "usual" point
Image

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: Updated GPU firmware

Fri Jan 18, 2013 12:45 pm

Is that after "pulling the plug" or a reboot? I'm finding that just unplugging it is not enough - I have to leave it powered off for a few minutes to have any any hope of getting it to boot. Always hangs at that "well known spot" if I try powering on too soon or with any overclocking/overvolting. I normally have it in a case along with a laptop drive and a pair of 4 port hubs (all powered from the same supply) - I removed it from the case and it doesn't seem to need the power off for quite as long which suggests (maybe?) there is something temp related going on. It normally runs fine for days, in the case, overclocked at 1100 about 55'-58'C without any of this CMA stuff...
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

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

Re: Updated GPU firmware

Fri Jan 18, 2013 1:11 pm

I applied oveclock and overvolt, run "shutdown -h now", unplugged, waited 120seconds, replug but it still hangs...I want to oveclock it....ho can I revert mi fwd to the latest version without CMA?

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: Updated GPU firmware

Fri Jan 18, 2013 1:30 pm

Just comment out all gpu_mem_256/512, cma_<stuff> in config.txt and put back whatever you might have had for gpu_mem= and you should be OK. I'm going to stay away from the CMA stuff unless/until something changes or I get some time to play around with a heatsink to prove/disprove there is a thermal component to this equation or I get some other goofy idea. In other words, the kernel/firmware seem fine, just leave CMA alone (for now). At least I haven't found a good reason to go back to the previous kernel/firmware.

Edit: I should have added that I needed to power off for about "n" minutes where"n" was greater than 2,
more like 5.
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

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

Re: Updated GPU firmware

Fri Jan 18, 2013 1:50 pm

Ok, thanks!

ricsi
Posts: 6
Joined: Sun Sep 02, 2012 12:09 pm

Re: Updated GPU firmware

Sun Jan 20, 2013 1:40 pm

I decided to try out CMA myself.

I have a RPi with 512MB RAM headless (set to 16 MB GPU) -> cannot gain anything so not checked.

I have anotherone with 256MB, that uses normally 32MB for GPU and I reboot with 64MB for Quake3.

So I added to the cmdline:
coherent_pool=6M smsc95xx.turbo_mode=N

And have in config.txt:
gpu_mem=64
#gpu_mem=112
cma_lwm=16
#cma_hwm=22
cma_hwm=32
cma_offline_start=16

I see this error in dmesg:
vc_cma_alloc_chunks: dma_alloc_from_contiguous failed for 40000 bytes (alloc 0 of 25, 25 free)

I think I have not seen it with gpu_mem=112
but I have also not seen it without CMA and GPU_mem 64.

PS: Any negative effects for avoid_pwm_pll=1 when I do not use the analog audio?!?
Want to decouple core_freq as it is not stable for my RPi on 500MHz

dom
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 5370
Joined: Wed Aug 17, 2011 7:41 pm
Location: Cambridge

Re: Updated GPU firmware

Sun Jan 20, 2013 6:56 pm

ricsi wrote: And have in config.txt:
gpu_mem=64
#gpu_mem=112

I see this error in dmesg:
vc_cma_alloc_chunks: dma_alloc_from_contiguous failed for 40000 bytes (alloc 0 of 25, 25 free)
You've reduced the size of CMA area (it comes from free gpu_mem), so not surprising that CMA allocs fail.
ricsi wrote: PS: Any negative effects for avoid_pwm_pll=1 when I do not use the analog audio?!?
Want to decouple core_freq as it is not stable for my RPi on 500MHz
Should be fine.

Return to “Advanced users”