runcom has quit [Read error: Connection reset by peer]
runcom has joined #u-boot
d-s-e has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
yollom_ is now known as yollom
ikarso has quit [Quit: Connection closed for inactivity]
persmule has quit [Remote host closed the connection]
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
d-s-e has quit [Ping timeout: 240 seconds]
mmu_man has joined #u-boot
prabhakarlad has quit [Quit: Client closed]
apritzel_ has joined #u-boot
slobodan has joined #u-boot
neha_ has joined #u-boot
<apalos>
sjg1: is there a reason we only open the device in tpm_init()?
<apalos>
because we could just call the newly added tpm_auto_start() which now handles both v1 and v2 and end up with a fully initialized TPM isntead of just claiming a locality
<apalos>
It would also make our life way simpler on testing
apritzel_ has quit [Ping timeout: 265 seconds]
d-s-e has joined #u-boot
<marex>
milkylainen: dtc -p 4096 goes somewhere
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
<marex>
rfs613: I reviewed as much as I could, I think it is mostly good to go
apritzel_ has joined #u-boot
<marex>
rfs613: I think there are a few items which can be fixed with sed, a memleak in mkimage and thats very much it
<milkylainen>
marex: pad? Not sure I follow. iiuc, the blob is expanded to make the new stuff fit when signing.
<milkylainen>
-S?
<milkylainen>
I'm just annoyed that it throws a whine when it's most likely going to succeed with the alloc.
<marex>
milkylainen: try 'git grep -i dtc.*4096' in UBoot sources
goliath has quit [Remote host closed the connection]
goliath has joined #u-boot
apritzel_ has quit [Ping timeout: 268 seconds]
<milkylainen>
marex: mkimage does the dtc call. I don't. Also. It sounds totally backwards to pre-reserve space for something that can be reallocated without whine.
<milkylainen>
or rather the lib calls.
<milkylainen>
It doesn't matter much. It reallocates by itself and continues. It's just whining.
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
neha_ has quit [Quit: Connection closed for inactivity]
<austriancoder>
sjg1: I have a question about device_notify(..). I am upgrading our U-Boot fork right now. We have ftl (flash translation layer) support in our fork and I a seeing a problem.
<austriancoder>
ret = device_notify(dev, EVT_DM_POST_PROBE); returns -1 on the first call it device_probe() returns -1 too. The device is fully probed. The second ls call succeeds as there is no need to probe the device.
<austriancoder>
here is some debug output I have added
<milkylainen>
So. The previous complaint I had about mkimage creating somewhat bogus error messages to the user, even when it fixes ENOSPC by itself.
<milkylainen>
First I fixed the error messages. All fine. But then I changed the starting size_inc in the incremental for loop in fit_image.c
<milkylainen>
Not only did I have no complaints, but the signing process took significantly less time.
<milkylainen>
My final dtbs did not change in size.
<milkylainen>
What would be a problem with starting with a size_inc > 0? It's not like a modern signing machine can spare 32k extra or something more useful than 0.
<sjg1>
milkylainen: Yes it would be better to make plenty of space and then remove it afterwards. The only problem is we don't want an FDT much larger than needed
<milkylainen>
sjg1: But it doesn't affect final size? Atleast not when I tried it?
<milkylainen>
I don't know if the code does final truncation to whatever the real size ends at?
<sjg1>
milkylainen: Perhaps it reduces it at the end...? I was not aware of it doing that
<milkylainen>
Maybe. I haven't checked. I did a binary diff between a dtb starting at 32k size_inc and default 0.
<milkylainen>
sjg1: I was doing a dts -> dtb + signing. Not a resign. If that affects things.
<milkylainen>
I'm assuming that the 1k steps and the * 10 execution time needed is because mkimage does 1k * 10 and computes the signature 10 times?
<milkylainen>
I think this particular "larger" dtb needed +10k.
<milkylainen>
For the signing.
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
<milkylainen>
ah. my bad
<milkylainen>
It is increased. But it should be possible to truncate to header size at the end?
<marex>
Tartarus: obv it will be postrc2, but that should be fine
<Tartarus>
kettenis: Oh, and is the M2 support ready for me to merge or should it go to next?
camus has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
<kettenis>
Tartarus: I think it is ready to merge
<kettenis>
I have some upcoming changes to upload firmware to the PCIe XHCI controllers, but those may be a bit more controversial and should probably go into next
<milkylainen>
sjg1: as far as I can tell from the specs, it should be fine to truncate the tree to the offset of the strings plus the size of the string section, aligned 8 bytes?
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
goliath has quit [Quit: SIGSEGV]
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
ladis has quit [Quit: Leaving]
<marex>
Tartarus: what do you supply that px30 from ?
<marex>
I assume you have a px30 at hand
runcom has quit [Quit: Konversation terminated!]
<marex>
is there like a USB charger or does it have a dedicated supply ?
runcom has joined #u-boot
Laurence_b has joined #u-boot
apritzel_ has joined #u-boot
<Tartarus>
I have a rockpro64, not a px30, sorry
<Tartarus>
And that's just a regular 12v barrel
<kettenis>
why am I not surprised all those coverity defects are in ipv6 code?
<rfs613>
px30 mini evb also seems to use a 12V barrel, fwiw
<Tartarus>
kettenis: in their defense, most of the big things went in for -rc1 and... I think coverity just ate the report for that one :(
<Tartarus>
Since I had to re-run it with a new version of their tools
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Read error: Connection reset by peer]
runcom has joined #u-boot
pgreco has joined #u-boot
<marex>
Tartarus: ah ok
<marex>
rfs613: thanks
* marex
goes back to digging in dwc2 weirdnesses
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
<sjg1>
milkylainen: Yes, but there is a libfdt function that does it all for you
Laurence_b has quit [Quit: Client closed]
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
slobodan has quit [Ping timeout: 268 seconds]
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
* marex
bangs head into desk
<marex>
so ... STM32MP15xx DWC2 does support BC 1.2 charger detection, but only on FS PHY (i.e. the one on dedicated pins), not on the HS PHY (i.e. the one everyone uses because HS)
runcom has quit [Quit: Konversation terminated!]
<konradybcio[m]>
Short them :brain:
runcom has joined #u-boot
runcom has quit [Quit: Konversation terminated!]
runcom has joined #u-boot
GNUtoo has quit [Remote host closed the connection]
GNUtoo has joined #u-boot
runcom has quit [Read error: Connection reset by peer]