04:40 <Guest40> Hi guys, Is there a way to change aarch64 system control registers in OpenOCD? 'mrs', 'msr' have removed long time ago..

09:03 <tarekb> orelem: do you want me to propose another version of #6569

09:05 <PaulFertser> I think it's just unclear when exactly and why that might be needed and why just setting WORKAREASIZE to 0 is not enough.

09:07 <tarekb> when you set WORKAREAD to 0, stm32l4_write_block will just fail

09:07 <tarekb> and there no other function to perform the flash programming

09:08 <tarekb> stm32l4_write_block_without_loader will do the job, if stm32l4_write_block returns ERROR_TARGET_RESOURCE_NOT_AVAILABLE

09:09 <PaulFertser> tarekb: I haven't looked deep into this topic but it feels really counter-intuitive for someone who only dealt with older families of stm32.

09:10 <tarekb> correct, that's why I have proposed to add more intelligence in the cfg file, to manage this intuitively

09:10 <PaulFertser> tarekb: so why can't it use the loader if working area is available and not use it when not, automatically?

09:12 <tarekb> I'm cooking a patch to do this switch in the cfg file, so it checks the security state and set the WORKAREA to zero when it should not be accessed

09:12 <tarekb> managing this at driver level will solve only the programming issue

09:13 <tarekb> so setting workarea size to zero will hint the verify and other functions to not use the target algo version ...

09:22 <PaulFertser> tarekb: so his revert is to remove the custom command from the driver. When is that command more useful compared to just setting WORKAREASIZE to zero?

09:24 <tarekb> we want to to partially revert this commit, so we remove the command stm33l4x flashloader (enable|disable) but we keep stm32l4_write_block_without_loader to perform the programming when workarea is zero

09:25 <tarekb> Ps the function stm32l4_write_block_without_loader was introduced in the same old patch

09:33 <PaulFertser> Makes sense

09:49 <tarekb> I have updated the #6569

09:50 <tarekb> but we are still missing the logic of setting workarea to zero in STM32L5 and STM32U5 cfg files

09:50 <tarekb> working on them ;)

13:10 <elpee> Hi all, since commit ca9dcc86d7d9f465950a0670a8c7a852ae2ca798 (https://review.openocd.org/c/openocd/+/4036/), we've been noticing some instabilities in firmware loading.

13:12 <olerem> elpee: it was mergend back in 2017

13:12 <elpee> we're loading our firmware to an stm32l082cz target, but sometime loading the firmware just elicits an "Error finishing flash operation"

13:13 <elpee> Indeed we were still on OpenOCD v0.10.0

13:13 <olerem> elpee: is it reproducable with latest git version?

13:13 <elpee> Yes

13:13 <olerem> tarekb: you help is needed ^

13:13 <olerem> tarekb: i didn't noticed you wrote to me, spelling issues ;)

13:14 <olerem> elpee: you should talk to tarekb, if he is not online right now, then please write to the mailing list

13:15 <elpee> I will send him a message, thanks.

13:17 <tarekb> olerem: I always write you nick name reversed

13:18 <tarekb> writing Oleksij is easier

13:18 <tarekb> I haven't noticed the messages I was with Antonio in the phone ;)

13:19 <Haohmaru> i hope your phone is at least 1.5x1.5 meters big

13:20 <tarekb> hahaha

13:33 <tarekb> FTR, in ST fork we are using this change https://review.openocd.org/c/openocd/+/5598 to fix ST32L0 flash issues

13:33 <tarekb> and it works like a charm

13:34 <tarekb> I will dive into it soon