NishanthMenon changed the topic of #openocd to: this is the place to discuss all things OpenOCD | Logs: https://libera.irclog.whitequark.org/openocd/
emeb has quit [Quit: Leaving.]
flatmush has quit [Ping timeout: 250 seconds]
flatmush has joined #openocd
tsal_ has quit [Ping timeout: 244 seconds]
tsal has joined #openocd
boru` has joined #openocd
boru has quit [Killed (NickServ (GHOST command used by boru`))]
boru` is now known as boru
Guest38 has quit [Ping timeout: 256 seconds]
Guest38 has joined #openocd
Guest38 is now known as ConstellationLab
c4017 has quit [Remote host closed the connection]
c4017 has joined #openocd
<ConstellationLab> I'm working with a fairly new STM32 chip and have not been able to get openocd to lock/unlock (memory protect and unprotect) the flash, but can do so with the STM32CubeProgrammer application. I can describe in more detail the problem, but wondering if there is anyone here that could help me modify the source or the TCL script to make this work so
<ConstellationLab> I can ultimately submit a pull request for a patch.
<ConstellationLab> I think I have narrowed it down by using a Saleae Logic Analyzer to record all the SWD transactions between the STLINK-V3SET and the STM32WLEx chip. With the STM32CubeProgrammer, as expected, there is a series of writes to the FLASH_OPTKEYR and FLASH_KEYR registers to unlock access to the FLASH_CR register (address 0x58004010)
<ConstellationLab> With OpenOCD, the FLASH_OPTKEYR and FLASH_KEYR sequence looks correct, but I only recorded reads from the FLASH_CR register and no writes, which are required to change the flash protection bits.
<ConstellationLab> Please note that this chip is silicon revision 1.1, which is not yet officially supported, so I edited the /src/flash/nor/stm32l4x.c to add this revision and recompiled from source (see my not-yet-submitted for Gerrit review merge request for the exact change: https://sourceforge.net/p/openocd/code/merge-requests/20/ )
<ConstellationLab> I look forward to any feedback you might have, thank you in advance!
Hawk777 has joined #openocd
emeb has joined #openocd
emeb has quit [Client Quit]
emeb has joined #openocd
c4017 has quit [Read error: Connection reset by peer]
emeb has quit [Quit: Leaving.]
ericonr has quit [Ping timeout: 240 seconds]
ericonr has joined #openocd
<olerem> ConstellationLab: i would recommend to contact the develope who provided this patch: https://review.openocd.org/c/openocd/+/6129
<olerem> or even better, write to openocd-devel@lists.sourceforge.net and CC Tarek :)
nerozero has joined #openocd
Hawk777 has quit [Quit: Leaving.]
Haohmaru has joined #openocd
indy has quit [Remote host closed the connection]
indy has joined #openocd
<karlp> st's started using numbers instead of letters? what's th world coming too?
* karlp kinda wants to look at the errata list now :)
<Haohmaru> where?
<Haohmaru> btw, i saw openocd in one of their documents.. that's good, i wish more $vendors would start acting like openocd (as well as other tools) exist
<karlp> well, guessing that wl1.0 had soem fun errata if they've got a 1.1 silicon
<karlp> st contributes directly to openocd these days, if you've missed that
<Haohmaru> kewl
<karlp> really not convinced that this tatic const struct stm32l4_rev stm32_462_rev style is more legible that what it replaced.
<Haohmaru> i don't really eyeball ST too much
<Haohmaru> oh you mean the chip part names?
<Haohmaru> STM32<letter><somenumbers> <- the letter is gone now?
<karlp> no, this is a recent code change.
<karlp> silicon revs used to be y,z,a,b style, ConstellationLab reports that the newest wl part is "1.1" instead.
<Haohmaru> oh that
<karlp> speaking of tarek...
<karlp> hrm, not here right now.
<karlp> PaulFertser: hacking still has openocd.zylin, do I need new instructions to push a change?
<karlp> or remote: You need 'Create' rights to create new references.
<karlp> remote: User: karlp
<Haohmaru> they were changing some stuff these days, which i don't understand, the word "gerrit" and "gertty" were being mentioned
<karlp> yeah, gerrit is upgraded, and review.openocd.org is the new endpoint, but not sure what is/wasn't migrated, andhacking docs still use the old urls
<Haohmaru> mr PaulFertser is definately pimping up the website
tchebb has quit [Quit: ZNC - http://znc.in]
tchebb has joined #openocd
<PaulFertser> karlp: hm, where are you pushing to?
<PaulFertser> What ref exactly?
<PaulFertser> karlp: via https or ssh?
<karlp> error: failed to push some refs to 'ssh://openocd.zylin.com:29418/openocd.git'
<karlp> reviewssh://karlp@openocd.zylin.com:29418/openocd.git (push)
<karlp> that's the git remote
<karlp> can I just change the url and push again?
<PaulFertser> karlp: you can but what is the remote ref? I think the problem is that ssh key changed. Can you try "ssh -p 29418 karlp@openocd.zylin.com" to verify if my guess is correct?
<karlp> I added my new ssh key for this machine.
<PaulFertser> karlp: hm, I'd like to see full output from gerrit, my guess is not correct most likely.
<PaulFertser> karlp: you should be pushing to refs/for/master , what are you pushing to?
<karlp> git push review, with whatever git hooks got setup in the past for that to all do the right thing?
<karlp> remote.review.push=HEAD:refs/publish/master
<karlp> is what's there from the past, has that changed?
<karlp> a6f5a167053fb2602f98671815f79d3382412d62 chagned it
<karlp> apparently gerrit deprecated that and it's now an error I guess.
<PaulFertser> karlp: yes, just use for/master instead, you can edit it in .git/config
<PaulFertser> Or review/master
<karlp> what the...
<PaulFertser> I think for/ is correct
<karlp> nevermind, it got pushed now, but it conclicts with more of tarek's stuff
<karlp> why is there device id patches since march that haven't been merged?
<PaulFertser> karlp: conflicts as shown by Gerrit are not too accurate.
<PaulFertser> karlp: Tarek's?
<PaulFertser> Or yours?
<karlp> the conflict shown is correct, tarek filed it in march it's n ot merged.
<karlp> also, "back" in gerrit doesn't work very well :)
<PaulFertser> Hm, I thought back should work, how to reproduce?
<PaulFertser> Not sure, Tarek is working hard on bringing upstream OpenOCD fully up-to-date regarding ST parts, probably just an oversight.
<karlp> I guess it's because it's part of a long sequence he has on the l4 flash driver.
<karlp> why is he using his gmail address? this is all commercial work isn't it?
<PaulFertser> karlp: why not? I am not using corporate e-mail for anything either.
<PaulFertser> I'm sure he spends some of his spare time too.
<PaulFertser> Probably a lot actually :)
<karlp> well like https://review.openocd.org/c/openocd/+/6109/18/contrib/loaders/flash/stm32/stm32l4x.c is copyright tarek@st, but filed by tarek@gmail...
<PaulFertser> Guess that's all right?
<PaulFertser> Copyright is one thing, actual usable e-mail address is another.
<Haohmaru> oh, so he's from ST
<karlp> ConstellationLab: you were just seeing "unknown" as revision right? you weren't running into any problems?
<PaulFertser> Antonio too
<Haohmaru> oh? that explains his involvement around the stlink3 "issue"
<Haohmaru> (i was wondering why he cares that much)
<PaulFertser> Antonio is volunteering _a damn lot_ , it's not just about ST work.
<Haohmaru> yeah, i saw
<Haohmaru> i eyeballed thru the sourceforge openocd codez the other day
ConstellationLab has quit [Ping timeout: 256 seconds]
gnom has quit [Ping timeout: 252 seconds]
gnom has joined #openocd
ntfreak has joined #openocd
cp- has joined #openocd
akaWolf has quit [Ping timeout: 244 seconds]
Guest38 has joined #openocd
<olerem> borneoa_: should I take https://review.openocd.org/q/topic:for_jim0.81 ?
Guest38 is now known as ConstellationLab
<ConstellationLab> @karlp Yes, before I made that change, chip revision 1.1 was just "unknown" so it wouldn't perform any function until I added the new one. After adding the 1.1 revision I could successfully program the chip, but locking and unlocking had to be done with ST tools.
<ConstellationLab> olerem: thanks for the tip, I'll write the the developer list and copy Tarek, thank you!
<karlp> ConstellationLab: it... should have worked, just printed unknown on the diags?
<karlp> the revision string is only ever used in prints, it doesnt gate any functionality, at least in the current code I'm looking at?
emeb has joined #openocd
akaWolf has joined #openocd
ntfreak has quit [Quit: Client closed]
ericonr has quit [Ping timeout: 252 seconds]
ericonr has joined #openocd
Haohmaru has quit [Remote host closed the connection]
nerozero has quit [Ping timeout: 244 seconds]
ConstellationLab has quit [Ping timeout: 256 seconds]
<karlp> so. my git submodule sitll points to the old jimtcl url.
<karlp> and .gitmodules has the right url, how can I convince it to update?
<karlp> git submodule sync --recursive apparently.
<karlp> thanks!
<karlp> damn autotootls, checkout a pathc, entire tree gets rebuilt :|
gnom has quit [Ping timeout: 252 seconds]
<karlp> meh, itm is all busted. timeout waiting for ITM_TCR_BUSY_BIT
<PaulFertser> On this new stm32?
<PaulFertser> Probably needs some special enablement as on other parts.
<karlp> no, works with 0.11.
* karlp is not looking at this anymore. enough to deal with this target at the moment.
gnom has joined #openocd
rue_mohr has quit [Remote host closed the connection]
gnom has quit [Ping timeout: 252 seconds]
gnom has joined #openocd
<karlp> 4gah, how do you turn off -Werror when building from a tag?
<PaulFertser> Just ./configure --disable-werror karlp ?
sbach has quit [Read error: Connection reset by peer]
sbach has joined #openocd
Helmholtz has quit [Changing host]
Helmholtz has joined #openocd
Guest38 has joined #openocd
Guest38 is now known as ConstellationLab
ConstellationLab has quit [Client Quit]
Fleck has quit [Remote host closed the connection]
Guest38 has joined #openocd
Guest38 has quit [Client Quit]
tarekb has joined #openocd
tsal has quit [Ping timeout: 240 seconds]
tsal has joined #openocd
tarekb has quit [Read error: Connection reset by peer]
<diddly> hi, i updated openocd from 0.10 to 0.11 in order to properly recognize this stm32l152 chip. now openocd is only listening on 127.0.0.1. is there a way to change the default back to 0.0.0.0?
<karlp> PaulFertser: probably would have worked.
<karlp> diddly: add "bindto 0.0.0.0
<karlp> " to your cfg file
<karlp> (or whatever you want to bind too)
<diddly> ok, but is there a system wide one? i have many config files for various boards (not the end of the world, just would prefer the old behaviour)
<karlp> hooray. security says no.
<karlp> this is what hapens when people do dumb shit and go cve shopping.
<karlp> you'd have to patch it out.
<diddly> oh it's almost definitely a better default, i'm just old and stuck in my ways, and very behind a firewall... all good :) the bindto did the trick as well, thank you very much
<karlp> yeah, tedious I know, but so be it...
emeb has quit [Quit: Leaving.]