NishanthMenon changed the topic of #openocd to: this is the place to discuss all things OpenOCD | Logs: https://libera.irclog.whitequark.org/openocd/
tsal has quit [Ping timeout: 240 seconds]
tsal has joined #openocd
thinkfat_ has joined #openocd
thinkfat has quit [Ping timeout: 272 seconds]
nerozero has joined #openocd
memes has quit [Read error: Connection reset by peer]
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
rkta has quit [Remote host closed the connection]
rkta has joined #openocd
Hawk777 has quit [Quit: Leaving.]
indy has quit [Quit: ZNC 1.8.2 - https://znc.in]
indy has joined #openocd
erhankur_ has joined #openocd
erhankur has quit [Read error: Connection reset by peer]
Haohmaru has joined #openocd
Bugies has joined #openocd
Bertl_zZ is now known as Bertl
erhankur_ has quit [Remote host closed the connection]
erhankur has joined #openocd
wingsorc__ has quit [Ping timeout: 244 seconds]
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Ping timeout: 252 seconds]
Bugies has quit [Read error: Connection reset by peer]
defiant has joined #openocd
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
Bugies has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
doelie has joined #openocd
<doelie> hi. question regarding rtt. my rtt buffer on the device is 64 bytes and i send 519 bytes over tcp. i see only a 63 byte write appear on the target. prob just my bug but just in case: does openocd drop data from tcp if it doesn't fit in the rtt buffer?
<karlp> that's a _realllly_ small rtt buffer aiui.
<doelie> it works with 1024
<karlp> I've not used it in anger mind you, only read the docs a bit.
<karlp> if you said 1500 does it work with 1024?
<karlp> if you _send_ sorry
<doelie> haven't tried, not easy to change atm... my max chunck size is 519. that works fine if i set buffer to 1024. so made me wonder if it silently drops data from tcp instead of buffering.
<doelie> happy to look at the code too... i found some rtt code just not the part yet where tcp is handled.
<doelie> rtt_server.c
<doelie> if i skim-read this correctly, rtt_input() reads all available bytes from the tcp connection via connection_read() then passes it on to rtt_write_channel()
<doelie> i think this then ends up in write_to_channel() in rtt.c
<doelie> and any excess seems to be silently dropped there in the MIN() operation
<doelie> doesn't seem straightforward to add backpressure such that it will only read enough data from tcp to not overflow the buffer... the place to do this seems to be rtt_input(): to there somehow get the available size in the rtt channel and instead of reading 1024 bytes, to read a limited amount.
<doelie> can't quite see where rtt_input is called..
<doelie> i guess this is polled somewhere (a service_driver?)
<doelie> ok that service->input() is called from the select loop... doesn't seem straightforward to leave the data in the socket buffer...
<doelie> karlp: thanks for pointing in the right direction, i was stuck in a wrong assumption about how this works
erhankur has quit [Remote host closed the connection]
<doelie> for my use i can work around it and use a larger buffer for now, though would be nice to support backpressure in openocd itself. i find the rtt facility quite useful.
<karlp> I wonder if rtt has protocol level methods for doing so? sounds kinda awkward to put it, but definitely buggy.
erhankur has joined #openocd
<doelie> karlp: i don't know. i don't think it's necessary since openocd could poll the size, but from the looks of it this doesn't really fit in the openocd service architecture. maybe simpler to explictly buffer the data somewhere else.
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
defiant has quit [Quit: defiant]
michalkotyla has quit [Quit: michalkotyla]
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Ping timeout: 240 seconds]
ikskuh has joined #openocd
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
crabbedhaloablut has quit [Remote host closed the connection]
crabbedhaloablut has joined #openocd
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
abby has left #openocd [#openocd]
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
shibboleth has joined #openocd
<shibboleth> PaulFertser, ping
erhankur has quit [Remote host closed the connection]
<PaulFertser> shibboleth: pong
erhankur has joined #openocd
<Haohmaru> table tennis time?
<shibboleth> hi, is there a specific reason why openocd-spi functionality hasn't been included in mainline/master?
<Xogium> :D
<PaulFertser> shibboleth: what's openocd-spi ?
<shibboleth> basically, more reliable/faster jtag/swd using rpi spi instead of "dumb" gpio butbang
<shibboleth> bitbang
<PaulFertser> shibboleth: interesting. I do not remember seeing those patches offered upstream.
<Haohmaru> hm
<Haohmaru> how fast does it go in SWD mode
<shibboleth> well, reason i'm asking is to figure out if there are dealbreakers/no-starters or simply that no one ever suggested merging it
<Haohmaru> possibly folks here weren't aware of it
<shibboleth> yeah, not discoverable by github search for some reason
<Haohmaru> thank M$ for that maybe
<shibboleth> that fork hasn't been updated in a while, you'll need https://github.com/lupyuen/openocd-spi/compare/master...Zorvalt:fix-multiple-gcc-10-errors.patch for clean compile on gcc10
Haohmaru has quit []
<PaulFertser> shibboleth: probably the author just never tried to send patches upstream.
<karlp> wow, issues a jtag-swd reset after every read, but that's ok, because it's now using spi to send the bits in bulk...
<karlp> nice trick
<shibboleth> you (maybe) remember/know me, mail and git issues. if anyone felt like submitting a pr for https://github.com/openocd-org/openocd/compare/master...lupyuen:master.patch and https://github.com/lupyuen/openocd-spi/compare/master...Zorvalt:fix-multiple-gcc-10-errors.patch it'd be great
<PaulFertser> shibboleth: OpenOCD expects contributions via Gerrit, not mail.
<shibboleth> even more complications. great :)
<shibboleth> maybe that last patch isn't required if the first is applied against an updated master
<PaulFertser> I'm not doing any active work on OpenOCD so I can't be the person caring about upstreming this work.
<karlp> they even say in their write up that the method for bolting it in is pretty nasty, so don't really see it being "just merge it yolo"
<karlp> someone's going to have to actually do the work.
<PaulFertser> Yes, I do not expect upstreaming to be trivial.
<shibboleth> well, at least i've mentioned it
<SteveMarple> Has anyone tested openocd-spi to see how the speed compares? I've come across that (or similar) previously but not tried it yet.
abby has joined #openocd
<shibboleth> well, i have some firmware modification of an electric scooter on my todo-list. haven't gotten to it and compared speed/reliability yet but happened upon openocd-spi during inital research
<shibboleth> initial
abby has left #openocd [#openocd]
<shibboleth> also, other than price, pro/cons between getting a segger jlink or some stlink?
<shibboleth> also also, anyone ever tested openocd with a buspirate?
<PaulFertser> jlink works with any target. stlink v3 works only with ST targets, and v2 is limited to SWD essentially, not good for JTAG chains.
<PaulFertser> Buspirate works but slow.
<shibboleth> 10-4
<shibboleth> i read something, something about jlink and updated profiles. will i have to register it to download new fw/settings/profiles?
<PaulFertser> I think OpenOCD doesn't use any advanced features of jlink so profiles do not come into play.
<shibboleth> but for use of stock software etc, registration required?
<PaulFertser> No idea about jlink software.
<shibboleth> k
<karlp> SteveMarple: somewhat surprisingly I can't find any speed reports in the author's blog posting other than "spi at 31MHz"
<karlp> none of the flashing examples have any time or even the gdb load speed included :|
<SteveMarple> I guess 31MHz is raw SWCLK frequency. With bcm2835 driver I see a bit over 1MHz. I've no feeling for how much those jtag-swd resets will cost in performance.
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
<karlp> hard to say, from the authors own use case it's clearly using far less host cpu at least :)
<shibboleth> i can do some testing, but eh... never gotten my feet wet with re to jtag before now, don't wanna mess up testing on the scooter. can anyone suggest something common/often laying around i can do some testing on?
<shibboleth> i've got some wd hdd pcbs, but tiny leads/contacts, something else
<karlp> it's not going to matter unless someone's interested in working on upstreaming it...
<shibboleth> also: in general
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
<shibboleth> karlp, also, i'd ofc do the testing running the fork vs running mainline/master on an rpi. merging it and then testing would be overkill for now :P
Hawk777 has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
Bertl is now known as Bertl_oO
nerozero has quit [Ping timeout: 255 seconds]
shibboleth has quit [Quit: shibboleth]
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
thinkfat_ has quit [Ping timeout: 244 seconds]
thinkfat_ has joined #openocd
Bugies has quit [Ping timeout: 258 seconds]
marex has quit [Ping timeout: 256 seconds]
wingsorc__ has joined #openocd
sbach has quit [Read error: Connection reset by peer]
sbach has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
akaWolf has quit [Ping timeout: 244 seconds]
akaWolf has joined #openocd
tsal has quit [Ping timeout: 240 seconds]
tsal has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd
erhankur has quit [Remote host closed the connection]
erhankur has joined #openocd