<zmatt>
rcn-ee: the shared ram really ought not be mapped as device memory in the first place, but I'm not sure how to fix that... maybe I should just use TI's cmem driver to allocate shared memory instead... assuming that's available on arm64 too?
<amir36>
different topic, all suppliers that I've checked only have A1 revision of BBAI, is there a way to buy the A2 revision? it has important modifications
<rcn-ee>
Zmatt, ti dropped cmem.. they are working on mainlining a dmabuf patch to support any remoteproc ti cores..
<rcn-ee>
It would allow us to make sure it's in the 32bit access range
<zmatt>
I'm pretty sure the allocation uio-pruss does is also in the 32-bit range
<zmatt>
anyway, I'll need to dig into what the options are... having to hardcode a reserved memory region in the DT seems pretty gross to me, and I'm not even sure how you'd map that to userspace. in the past I've always used uio for that
<rcn-ee>
That'll be the next thing.. someone will want access to that data thru dev mem from the arm side..
<zmatt>
/dev/mem should never be needed
<zmatt>
I think the problem should also be fixable by changing the memtype of the ddr memory region in uio_pruss.c from UIO_MEM_PHYS to UIO_MEM_IOVA ... though that would be specific to devices with coherent memory such as the j721e
<zmatt>
so that would either be a slight difference between uio-pruss in the arm64 kernel vs the 32-bit one, or it would need a #if
<zmatt>
I can probably also work around the problem in py-uio, and might need to do that anyway since changing the memtype isn't really an option for the pruss-local memories
<zmatt>
though changing the memtype for the ddr memory would also be a good idea for performance
amir36 has quit [Quit: Client closed]
<zmatt>
ideally I'd want memory mapped registers to use whatever memory type the linux kernel uses for that same purpose (device memory on arm, presumably device-nGnRE on arm64 though I'd need to check), normal non-cacheable for non-coherent memory (e.g. pruss memory, or all memory on non-coherent SoCs like the am335x and am572x), and normal cacheable on coherent SoCs like the j721
<rcn-ee>
do you want to add a v3 compat option to help define teh difference? .compatible = "ti,pruss-v3" or just assume aarch64 is that way?
amir36 has joined #beagle
<zmatt>
for now I think just assuming aarch64 = coherent would be fine I think, since that is currently true for any aarch64 device that has pruss... for a cleaner solution I'd probably want something better than a different compatible-string since it's not really about the pruss, it's about the architecture
<amir36>
guys and gals, as crazy as it sounds, I created a post on the forum (pending now), inviting people to collaborate on making embedded development as easy and as well documented as high-level programming. I really hope that we come together and educate each other and newer engineers. If my post got approved, please show your support if you also think
<amir36>
programming our computers should not be as complicated as it is now. I think BB SBCs are a great place to start.
<zmatt>
rcn-ee: uio should probably use phys_mem_access_prot() just like /dev/mem does
thinkfat has quit [Ping timeout: 240 seconds]
thinkfat has joined #beagle
starblue has quit [Ping timeout: 276 seconds]
starblue has joined #beagle
amir36 has quit [Quit: Client closed]
brook has joined #beagle
vagrantc has quit [Quit: leaving]
Shadyman has joined #beagle
set_ has joined #beagle
brook_ has joined #beagle
buzzmarshall has quit [Quit: Konversation terminated!]
brook has quit [Ping timeout: 240 seconds]
brook_ has quit [Ping timeout: 240 seconds]
Guest42 has quit [Ping timeout: 252 seconds]
Guest42 has joined #beagle
akaWolf has quit [Ping timeout: 256 seconds]
akaWolf has joined #beagle
akaWolf has quit [Ping timeout: 244 seconds]
akaWolf has joined #beagle
k-man has quit [Ping timeout: 268 seconds]
ikarso has joined #beagle
otisolsen70 has joined #beagle
demirok has joined #beagle
bradfa has quit [Read error: Connection reset by peer]
denix has quit [Ping timeout: 244 seconds]
pbrobinson has quit [Ping timeout: 244 seconds]
bradfa has joined #beagle
vitorio has quit [Read error: Connection reset by peer]
mturquette has quit [Read error: Connection reset by peer]
denix has joined #beagle
vitorio has joined #beagle
mturquette has joined #beagle
pbrobinson has joined #beagle
Guest42 has quit [Ping timeout: 252 seconds]
willow has quit [Ping timeout: 240 seconds]
willow__ has joined #beagle
Steve_ has joined #beagle
SJFriedl has quit [Ping timeout: 264 seconds]
ikarso has quit [Quit: Connection closed for inactivity]
indigaz has joined #beagle
indigaz8 has quit [Ping timeout: 276 seconds]
willow__ is now known as roughnecks
starblue has quit [Ping timeout: 240 seconds]
starblue has joined #beagle
ikarso has joined #beagle
<set_>
@zmatt: When you have time, will you please review this 2020 source for the ServoCape? https://pastebin.com/cxkVJ8Nw
<set_>
For some reason, unbeknown to me, I have a resource busy. does config-pin work in this matter or does an overlay need to be added?
<set_>
Anyway...I guess things are taking place now. So, I will stay patient. I think the i2c bus is open but busy. So, I will just wait until things get arranged.
roughnecks has quit [Killed (NickServ (GHOST command used by roughnecks_!~irc@user/roughnecks))]
brook has joined #beagle
Guest4458 has joined #beagle
xet7 has joined #beagle
Guest4458 is now known as puke
puke is now known as Guest9327
Guest9327 has quit [Killed (lithium.libera.chat (Nickname regained by services))]
Guest9327 has joined #beagle
Guest9327 has quit [Remote host closed the connection]
trifling has joined #beagle
set_ has quit [Read error: Connection reset by peer]
set_ has joined #beagle
johanhenselmans has quit [Read error: Connection reset by peer]
johanhenselmans has joined #beagle
Shadyman has quit [Quit: Leaving.]
brook_ has joined #beagle
brook has quit [Ping timeout: 260 seconds]
brook_ has quit [Ping timeout: 255 seconds]
mag has quit [Ping timeout: 276 seconds]
brook has joined #beagle
demirok has quit [Quit: Leaving.]
agrue has joined #beagle
mag_ has joined #beagle
buzzmarshall has joined #beagle
vagrantc has joined #beagle
Guest19 has joined #beagle
Guest19 has quit [Client Quit]
amir36 has joined #beagle
Guest42 has joined #beagle
amir_ has joined #beagle
amir_ is now known as Guest5806
Guest5806 has quit [Remote host closed the connection]
Guest5806 has joined #beagle
amir36 has quit [Ping timeout: 252 seconds]
Guest5806 is now known as Amir123
brook has quit [Quit: Leaving...]
mattb00ne has quit [Ping timeout: 264 seconds]
mattb0ne has quit [Ping timeout: 264 seconds]
brook has joined #beagle
set_ has quit [Remote host closed the connection]
set_ has joined #beagle
Amir123 has quit [Remote host closed the connection]
Amir123 has joined #beagle
ikarso has quit [Quit: Connection closed for inactivity]
Guest42 has quit [Quit: Client closed]
xet7 has quit [Quit: Leaving]
akaWolf has quit [Ping timeout: 240 seconds]
Amir123 has quit [Ping timeout: 244 seconds]
Amir123 has joined #beagle
akaWolf has joined #beagle
EdwardCromwell has joined #beagle
EdwardCromwell has quit [Client Quit]
Amir123 has quit [Remote host closed the connection]
Amir123 has joined #beagle
waldo323_ has joined #beagle
waldo323 has quit [Read error: Connection reset by peer]
waldo323__ has joined #beagle
waldo323__ has quit [Remote host closed the connection]
waldo323_ has quit [Ping timeout: 240 seconds]
ikarso has joined #beagle
<zmatt>
set_: check the kernel log for messages, I think that error generally means a problem with the i2c bus
<zmatt>
also, that address looks wrong
<set_>
Okay. Thank you. I just checked back and blamo. A message!
<zmatt>
you're passing 0b1000000 while the correct value based on the schematic is 0b111111
<set_>
Dang it.
<set_>
Oh!
<set_>
I broke the machine earlier. Something about an overlay and the wrong kernel.
<set_>
Anyway, off to try another route. This pup is not booting.
<set_>
Once I get back to where I was, before the crash, I will check the kernel log for messages.
v0n has quit [Quit: WeeChat 3.5]
vvn has quit [Quit: WeeChat 3.5]
vvn has joined #beagle
v0n has joined #beagle
<zmatt>
oh, this thing does have an overlay
<zmatt>
although it looks a bit broken, I don't understand why it's declaring the inputs as keys (and that declaration is in fact also broken)
<set_>
Right. But...
<set_>
Right.
<set_>
Outside of that idea, I just found a file. It is called and I quote, "My First Tensor."
<set_>
Poots.
<zmatt>
though more importantly, if you use the overlay then you can't talk to the pca9685 directly since it'll be in use by the kernel driver
<set_>
Aw.
<zmatt>
though that declaration looks broken too
<set_>
Aw!
<zmatt>
the whole overlay is full of typos
<set_>
Hmm.
<zmatt>
it's also messing with i2c2, which it shouldn't
<set_>
The overlay?
<set_>
Hmm. That is probably the reason.
<zmatt>
I suggest removing the BBORG_SERVO-00A2 overlay
<set_>
Right. I almost had the chance.
<set_>
it broke my system.
<set_>
I will not add it this time.
<zmatt>
add it? pretty sure it's installed by default?
<zmatt>
at least, it's part of the bb.org-overlays repository
<set_>
Oh no...I understand. I did not concoct it myself but I was...using BeagleBone-DeviceTrees instead.
<set_>
Not the ...bb.org-overlays.
v0n has quit [Quit: WeeChat 3.5]
vvn has quit [Quit: WeeChat 3.5]
vvn has joined #beagle
agrue has quit [Quit: Ping timeout (120 seconds)]
v0n has joined #beagle
<set_>
Buggers!
zjason has quit [Read error: Connection reset by peer]
<set_>
Whelp...I done looked around. I think GPIO-OF-HELPER is causing the issue but who knows?
<set_>
dang it. brb.
<zmatt>
ehm, no
zjason has joined #beagle
<set_>
No?
<set_>
It allocates a bunch of GPIO pins...I am using i2c2, I thought.
<set_>
Argh. okay. I will use i2c1.
<zmatt>
i2c2 is the correct bus, and is configured to i2c mode by default
<set_>
Oh.
<set_>
Okay.
<set_>
Aw!
<set_>
There is a timeout.
<set_>
omap_i2c 4819c000.i2c: timeout waiting for bus ready
v0n has quit [Quit: WeeChat 3.5]
vvn has quit [Quit: WeeChat 3.5]
<set_>
PRU is available!
<set_>
CAN is up.
<zmatt>
"timeout waiting for bus ready" means either no cape is connected or something is messing up the bus
vvn has joined #beagle
<set_>
Hmm.
<set_>
Stinks.
<set_>
IRQ?
otisolsen70 has quit [Quit: Leaving]
<zmatt>
no as in, external hardware
<set_>
Oh.
<zmatt>
either that or you changed the pinmux of the pins
<set_>
I did.
<zmatt>
but it should be correct by default
<zmatt>
well then whatever you did is probably the porblem
<set_>
I used config-pin p9.21 i2c
<zmatt>
*problem
<set_>
p9.22 i2c
<set_>
Oh.
<zmatt>
those are the wrong pins, and doing that messed up the bus
<set_>
Well. I will put them back to default.
<set_>
Still. Device or resource busy.
<set_>
One thing I noticed that did change...
<set_>
PM: Bootloader does not support rtc-only!
<zmatt>
I have no idea what that means but it's irrelevant
<set_>
and my CAN devices are registered now.
<set_>
Oh. Okay.
<zmatt>
and did you disable/remove the servo cape overlay?
<set_>
I rewrote the entire image. Yes, the overlay is nowhere to be found on my system but...
demirok has joined #beagle
<set_>
u-boot detected the cape.
<zmatt>
I mean, of course it does... that's why you need to ensure the cape is either removed or explicitly disabled
<zmatt>
the overlay I mean
<zmatt>
sorry
<set_>
I understand.
<zmatt>
anyway, afk
<set_>
The Cape is on and there is no overlay except dmesg is showing uboot recongnizing it.
<set_>
Oh.
<set_>
Okay. No issue. After awhile!
Posterdati has quit [Remote host closed the connection]
Amir123 has quit [Ping timeout: 244 seconds]
brook has quit [Remote host closed the connection]
brook has joined #beagle
demirok has quit [Quit: Leaving.]
marcheu has quit [Ping timeout: 244 seconds]
demirok has joined #beagle
marcheu has joined #beagle
brook has quit [Remote host closed the connection]
brook has joined #beagle
brook has quit [Remote host closed the connection]
brook has joined #beagle
brook has quit [Remote host closed the connection]
brook has joined #beagle
Shadyman has joined #beagle
v0n has joined #beagle
vagrantc has quit [Quit: leaving]
ikarso has quit [Quit: Connection closed for inactivity]
brook has quit [Remote host closed the connection]