<Seppoz>
in the sunxi 3.4 kernel i dont see under watchdog timer the suxi watchdog
<Seppoz>
how do i enable the kernel watchdog
deepe has joined #linux-sunxi
<Seppoz>
there is no earlyprintk in the uboot kernel args
<NiteHawk>
try "grep -i printk .config". you might have CONFIG_EARLY_PRINTK?
reinforce has quit [Quit: Leaving.]
<Seppoz>
nope
<Seppoz>
my bad was looking at the old .confg
<Seppoz>
found it thanks
<oliv3r>
Turl: ping
<oliv3r>
:p
<Turl>
oliv3r: pong
<Turl>
oliv3r: sup
cubear has joined #linux-sunxi
ssvb has joined #linux-sunxi
ricardocrudo has quit [Ping timeout: 246 seconds]
<ssvb>
Seppoz: pong
<Turl>
ohai ssvb
<Turl>
ssvb: /topic
<ssvb>
Turl: that's great, thanks for your hard work
<Turl>
ssvb: yw
<Turl>
ssvb: it's still running on the old server that's bad, that's why it's RO
<Seppoz>
ssvb: we had a capacitor for the uboot switch
<Seppoz>
ssvb: to debounce it.
<Turl>
at least you can read and point people to it until we get the new one set up
<Seppoz>
ssvb: the capacitor took to long to get the pin high
<Seppoz>
ssvb: which caused the board to boot in FEL instead of regular boot. i removed this capacitor and now it seems to allways boot fine
<Seppoz>
ssvb: thanks again
<ssvb>
Seppoz: you are welcome
<Seppoz>
still way better than having to worrie about reset and stuff :P
<ssvb>
are the reboot problems also resolved now?
Renard has quit [Remote host closed the connection]
<Seppoz>
the reboot problem i think is distro caused
<Seppoz>
i think it never shut down
<ssvb>
I think that shutdown definitely needs axp, but mripard_ may correct me
Renard has joined #linux-sunxi
<Seppoz>
the shutdown did work fine on a10 without axp
<Seppoz>
but ill takke a closer look on this later
<ssvb>
Turl: are we still waiting for the new server?
<Seppoz>
anyone know where autoboot_delay is defined in uboot 2014?
<Turl>
ssvb: we got it already, it just needs to be set up
tomcheng76 has quit [Read error: Connection reset by peer]
mmarker has quit [Read error: Connection reset by peer]
ricardocrudo has joined #linux-sunxi
tomcheng76 has joined #linux-sunxi
<Seppoz>
anyone know where CONFIG_BOOTDELAY is located in uboot 2014? in 2012 it was un sunxi_common.h. however its moved
<ssvb>
Seppoz: try to grep u-boot sources?
<Seppoz>
yea its 50 pages of results
<Seppoz>
seems like any other board have it in the configs header
cnxsoft has quit [Quit: cnxsoft]
gianMOD has quit [Remote host closed the connection]
tomcheng76 has quit [Read error: Connection reset by peer]
<ssvb>
Seppoz: ignoring the board specific configs, it is found in 'include/config_distro_defaults.h' and 'include/env_default.h'
<ssvb>
Seppoz: I guess that you need to either add it back to sunxi_common.h or maybe patch config_distro_defaults.h
<Seppoz>
i guess i can just change distro_default
<Seppoz>
thanks
tomcheng76 has joined #linux-sunxi
gianMOD has joined #linux-sunxi
gianMOD has quit [Ping timeout: 240 seconds]
reinforce has joined #linux-sunxi
orly_owl has quit [Quit: leaving]
<plaes>
\o/ wiki is up
<rellla>
Turl: thanks!
<Turl>
yw :)
<Turl>
keep in mind it's read only though
<Seppoz>
only my USB0 OTG seems to work, do i have to enable the other 2 in the driver somehow? i have them enabled in the fex and the devices are powered (5V are there)
<Turl>
Seppoz: did you enable the ohci/ehci host driver?
<Seppoz>
yes
<Seppoz>
CONFIG_USB_SUNXI_EHCI=y
<Seppoz>
CONFIG_USB_SUNXI_OHCI=y
<Turl>
_COMMON, _USB_MANAGER and what not too iirc
<ssvb>
nothing in particular, just a lot of usb stuff is compiled as modules, so it would not be happy if the modules are not installed correctly
<Seppoz>
they should be
<ssvb>
it might be a good idea to pay special attention to [usbc1] and [usbc2] sections in fex, and maybe borrow some stuff from the other fex files used on similar boards
<Seppoz>
this fex file allready works fine on a10
<Seppoz>
with older kernel tho
<Seppoz>
so ieither my kernel is misconfgured
<Seppoz>
or i require different settings with newer kernel
<Turl>
CONFIG_USB_OHCI_HCD_PLATFORM=y
<Turl>
CONFIG_USB_EHCI_HCD_PLATFORM=y
<Turl>
try disabling those two
<Seppoz>
ok
<Seppoz>
usb_restric_flag = 0
<Seppoz>
do i need that?
<Seppoz>
that must be new
leviathanch has joined #linux-sunxi
igordcard has joined #linux-sunxi
<Seppoz>
looks better now
<Seppoz>
thank you sir
igordcard has left #linux-sunxi ["Leaving"]
gianMOD has joined #linux-sunxi
leviathanch has quit [Ping timeout: 246 seconds]
leviathanch has joined #linux-sunxi
Andy-D has joined #linux-sunxi
gianMOD has quit [Ping timeout: 256 seconds]
mmarker has joined #linux-sunxi
leviathanch has quit [Remote host closed the connection]
domidumont has quit [Read error: Connection reset by peer]
FreezingAlt has quit [Quit: Out]
leviathanch has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
leviathanch has quit [Quit: No Ping reply in 180 seconds.]
leviathanch has joined #linux-sunxi
domidumont has joined #linux-sunxi
leviathanch has quit [Ping timeout: 246 seconds]
Guest78775 has joined #linux-sunxi
leviathanch has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 264 seconds]
vishnup has joined #linux-sunxi
gianMOD has joined #linux-sunxi
gianMOD has quit [Ping timeout: 272 seconds]
mmarker has quit [Read error: Connection reset by peer]
konradoo77 has joined #linux-sunxi
mmarker has joined #linux-sunxi
mmarker has quit [Changing host]
mmarker has joined #linux-sunxi
<Fsanches>
is there somewhere an overview of the purpose of each of the linux-sunxi git branches at github ?
aballier has quit [Quit: leaving]
<Fsanches>
if I want to submit patches, what's the branch my work should be based on?
Progman3K has joined #linux-sunxi
leviathanch has quit [Quit: No Ping reply in 180 seconds.]
<Progman3K>
How can I boot an aw-som10 from a flash-drive? I have the rootfs and other files on my card but how do I tell the board to boot from it rather than its MMC?
leviathanchcn has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
<WarheadsSE>
root= ?
mmarker has quit [Ping timeout: 264 seconds]
<vishnup>
ssvb:?
<ssvb>
vishnup: ...
Nyuutwo has quit [Quit: No Ping reply in 180 seconds.]
book`_ has quit [Ping timeout: 252 seconds]
alexxy has quit [Quit: No Ping reply in 180 seconds.]
<vishnup>
I got some time to try libdram, however, spl size increased to ~18KB
kz1 has quit [Read error: Connection reset by peer]
nashpa has quit [Ping timeout: 252 seconds]
Nyuutwo has joined #linux-sunxi
mcan has quit [Ping timeout: 252 seconds]
<vishnup>
can not be loaded to SRAM
alexxy has joined #linux-sunxi
lioka has quit [Ping timeout: 252 seconds]
sfritsch has quit [Ping timeout: 252 seconds]
kroylar has quit [Ping timeout: 252 seconds]
kz1 has joined #linux-sunxi
lioka has joined #linux-sunxi
sfritsch has joined #linux-sunxi
kroylar has joined #linux-sunxi
<vishnup>
BTW, thanks for FEL regression fix patch, I've tested on A33 tablet, now it
<vishnup>
starts SPL
book` has joined #linux-sunxi
nashpa has joined #linux-sunxi
mcan has joined #linux-sunxi
<vishnup>
ssvb: A33 has SRAM A1: 32KB SRAM A2: 64 KB, I think is possible to load > 18KB SPL, isn't it?
mcan has quit [Changing host]
mcan has joined #linux-sunxi
<ssvb>
vishnup: are you running out of available space?
<ssvb>
vishnup: a part of SRAM is used by the BROM FEL code and not available to us
<vishnup>
while writing SPL itself it gives error:
<vishnup>
usb_bulk_recv error -9
ricardocrudo has quit [Ping timeout: 255 seconds]
iamfrankenstein has joined #linux-sunxi
<NiteHawk>
can a 'monolithic' u-boot-sunxi-with-spl.bin be used in FEL mode, or it it SD card only?
<NiteHawk>
I figure that (due to size limitations) FEL would always require a reasonably small SPL?
<vishnup>
NiteHawk: u-boot-sunxi-with-spl.bin can not be used in FEL
<NiteHawk>
ok, thx for clarifying that. I'm okay with separate spl + u-boot
konradoo77 has quit [Ping timeout: 246 seconds]
<ssvb>
vishnup: wasn't it the dram code timing out for you earlier?
<ssvb>
vishnup: now you are saying that the SPL upload fails, I'm a little bit confused
<ssvb>
vishnup: or do you mean that linking libdram has increases the SPL size and it does not fit anymore?
<vishnup>
If I exclude libdram, it's able to upload and execute with DRAM timeout
<vishnup>
upload prob comes only if libdram is included.
gianMOD has joined #linux-sunxi
vishnup has quit [Remote host closed the connection]
vishnup has joined #linux-sunxi
Vespine has joined #linux-sunxi
<Progman3K>
How can I boot an aw-som10 from a flash-drive? I have the rootfs and other files on my card but how do I tell the board to boot from it rather than its MMC?
<vishnup>
ssvb: Is MAX size FEL can write is fixed?
wickwire has quit [Read error: Connection reset by peer]
cubear has quit [Ping timeout: 264 seconds]
<vishnup>
ssvb: I'll try more things tomorrow morning.
<Progman3K>
How can I boot an aw-som10 from a flash-drive? I have the rootfs and other files on my card but how do I tell the board to boot from it rather than its MMC?
<ssvb>
NiteHawk: for 2015.04
<Cooper_>
ssvb: Over the last 24 hours I've been running lima-memtester together with cpuburn-a7. Everything's rock-solid. No error reported, rotating block on the screen, gray background. I'm thinking I really did just over-rice my system.
<Progman3K>
All it says in the wiki for that is "Now you should be able to unmount your SDCard filesystems, and you should be able to boot your brand new installation."
<Progman3K>
That's not really helpful
sono has joined #linux-sunxi
gianMOD has quit [Ping timeout: 240 seconds]
<sono>
is there a banana* specific channel anyone is aware of?
<ssvb>
vishnup: yes, the SPL size is limited, but switching to Thumb2 saves some space and it might be enough to fit
<ssvb>
Cooper_: but it needs ruby interpreter, and also userspace cpufreq governor enabled in the kernel (which might be not there by default)
<Cooper_>
Oh, I see. For each frequency there's a different voltage and all I've tested thus-far is full throttle.
<ssvb>
Cooper_: it still might switch back and forth if there is some idle time
<Cooper_>
ssvb: idle time vs cpuburn-a7....
<ssvb>
Cooper_: during emerge, the cpu load is non-uniform, because there are configure and installation steps, which run on just a single core
<ssvb>
Cooper_: what kind of cpufreq governor are you using right now?
<Cooper_>
ssvb: Of course. I meant that thus far (=since running lima-tester along with cpuburn-a7, which is the time I've been verifying the build) there shouldn't have been idle time.
<Cooper_>
ssvb: ondemand
<ssvb>
Cooper_: for a non-battery powered board, the 'performance' governor may be a better choice
<Cooper_>
ssvb: But doesn't that keep the chip at max speed? I want performance when I'm home (=30% of the time) and lower power draw when I'm not in a position to use the devices (sleep, work, etc).
dack has joined #linux-sunxi
gianMOD has joined #linux-sunxi
<wens>
i don't think a couple of cortex-a7 cores will draw much power, even at full speed
<Cooper_>
wens: True. The difference between idle and full blast is 3W.
<ssvb>
Cooper_: we have some power consumption measurements, which show that the idle power consumption difference is negligible between the 'performance' and 'ondemand' governors on a20
<Cooper_>
ssvb: So it does scale down in some way?
ricardocrudo has quit [Read error: Connection reset by peer]
<ssvb>
Cooper_: the clock is probably disabled completely, not just reduced
<Cooper_>
ssvb: I see. I'll give that a go in a bit, see where she stands. I already have pretty good measurements thus-far, though they are from my complete setup rather than a single board, but the other ones are literally doing nothing so their power consumption should be constant (hence the "3W difference" quote - I did measure deviation from the norm).
<wens>
what board are you using?
<Cooper_>
wens: PcDuino3 Nano
<wens>
hmm
<Cooper_>
wens: That's an A20 with 1 GB ram.
<wens>
so that's without regulators dropping voltage
<ssvb>
Cooper_: the power consumption difference between idle and full load is naturally expected
<ssvb>
Cooper_: the question was whether idling at 60MHz is providing any reasonable power saving vs. idling at 1GHz
<Cooper_>
wens: As I said, I don't know idle power draw for a single board, but when I have 2 of those Nanos, a 12V-to-5V buck converter, 3 desktop SATA harddisks in low power state (drive motor off) and a 1-to-5 SATA port multiplier with those 3 harddisks attached but not doing anything, The total power draw of all that is just under 10W.
<ssvb>
Cooper_: and the tests have shown that the power consumption difference is negligible, but the responsiveness becomes pretty bad (cpufreq is too slow to increase the clock speed)
<Cooper_>
wens: Put cpuburn across both cores on just 1 of those Nanos and the total power consumption goes up to around 12W.
<Cooper_>
ssvb: As I said, I'll be trying it out as soon as my kernel module for the userspace cpufreq governor finishes compiling.
<wens>
so you got the port multiplier working?
<ssvb>
Cooper_: cpuburn is a bit artificial workload, you can realistically get somewhat close to it only when doing video transcoding or neon accelerated bitcoin mining :)
<wens>
we used to use seti@home for x86 load testing :p
<Cooper_>
wens: Yup.
<Cooper_>
ssvb: With performance the power draw difference of the idle machine is effectively 0. I'll stick with that for 'production'. Thanks.
adj_ has quit [Ping timeout: 245 seconds]
<ssvb>
wens: what do you think about the "race to idle" approach?
<ssvb>
wens: some people think that on modern processors reducing the cpu clock speed only makes sense for overheating prevention
<ssvb>
wens: and in the case of Allwinner chips, stressing the CPU with the 1.4V core voltage for prolonged periods of time may be not very nice
<wens>
race to idle?
<ssvb>
wens: the idea that it is best to complete the task as fast as possible (ex. by running at 1GHz) and go to sleep sooner than crawling at a lower clock speed
<wens>
i think it makes sense
<wens>
i mean the idle core can go into deeper sleep states
<Turl>
NiteHawk: it's probably not usable, so it's moot
<Turl>
for the most part, you can ignore anything that's not 3.4 on that repo
<NiteHawk>
oh, ok - so sunxi-3.4 is still "state of the art"?
<Turl>
NiteHawk: if you want full hardware support, yes
konradoo77 has quit [Ping timeout: 264 seconds]
iamfrankenstein has joined #linux-sunxi
<Turl>
NiteHawk: which these days mostly means display, 2D accel and mali
<Turl>
if you can live without those you can use mainline
<NiteHawk>
i see
konradoo77 has joined #linux-sunxi
adj_ has joined #linux-sunxi
tomcheng76 has quit [Read error: Connection reset by peer]
Vespine has quit [Quit: Leaving]
tomcheng76 has joined #linux-sunxi
interrobangd has joined #linux-sunxi
<interrobangd>
its online again!
<interrobangd>
ehx.
<interrobangd>
*thx
<interrobangd>
.. hetzner :|
FlibberTGibbet has quit [Quit: Leaving]
<Seppoz>
whats hetzner?
<interrobangd>
hoster
<Seppoz>
ah
konradoo77 has quit [Ping timeout: 264 seconds]
gianMOD has quit [Remote host closed the connection]
<Cooper_>
ssvb: Tried the cpufreq-ljt-stress-test and it completes immediately saying it's passed the test. I'm not getting the output with various MHz reported as the page does and as I said, it completes pretty much immediately. It's hard to believe it's done anything.
<Cooper_>
c/does/shows/
<ssvb>
Cooper_: does it show any error messages or something?
<Cooper_>
ssvb: Does it matter that I made that governor a module and not the default just yet?
<ssvb>
Cooper_: maybe try to "modprobe cpufreq_userspace" to see if it helps?
<Cooper_>
ssvb: Already loaded. From lsmod: cpufreq_userspace 2585 1
dack has quit [Remote host closed the connection]
akaizen_ has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
<ssvb>
Cooper_: google says that some people also had 'scaling_available_frequencies' missing, but I'm still trying to figure out why this could be happening
pmattern has quit [Quit: Genug für heute.]
<Cooper_>
ssvb: Building the kernel with it built-in rather than as a module.
akaizen has quit [Ping timeout: 272 seconds]
ricardocrudo has quit [Ping timeout: 264 seconds]
<Cooper_>
ssvb: Still nothing...
fr4z has quit [Remote host closed the connection]
viccuad has joined #linux-sunxi
<Turl>
interrobangd: yeah, it's the old server
<Turl>
that's why it's read-only
<interrobangd>
with new disk
<Turl>
interrobangd: with old disks
<interrobangd>
but i thought its dead and hetzner has changed it
<Turl>
we replaced one of the disks, but there's others that are bad too
<Turl>
so we're moving to a new server with new disks
<interrobangd>
but only the disks are bad, everyting else is still ok?
* interrobangd
is happy to see the wiki again
<Turl>
well, it makes you think if there's nothing else bad on there
<Turl>
iirc the PSU died some time ago too
<Turl>
and was replaced
<Cooper_>
ssvb: Isn't there a way to just outright tell that stresstester the frequencies it could attempt which it could verify against the min and max values reported?
<ssvb>
Cooper_: you can also enable cpufreq stats in the kernel, the script can also extract the frequencies there
<ssvb>
Cooper_: but it is quite an interesting case, I would need to reproduce the problem
<Cooper_>
ssvb: Looked at the script. It's the time_in_state file you're now referring to, yes?
<ssvb>
Cooper_: yes
<Turl>
is CPU_FREQ_TABLE enabled?
<Cooper_>
ssvb: Is that freq_list array populated with strings or numbers? I'm going to hard-code it for now based on the numbers on the hardware reliability test page.