Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = = ...quot; to stm32.cfg
David Brownell davi...@pacbell.net napisał(a): I think Freddie's comment (that it does have something to do with it) made confusion, then. Which comment? That would be explained by my trusting Freddie's comment to be accurate, i while it was instead ncorrect at the levels I was commenting on. ? 4\/3!! ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 02/12/2010 13:15, Spencer Oliver wrote: On 02/12/2010 12:42, freddie_cho...@op.pl wrote: Spencer Olivers...@spen-soft.co.uk napisał(a): As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - this has undesired effects for people who expect srst for example to work. We need to change that functionality then adding your patch would be ok. This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. I am looking at this at the moment, as i would like a solution. Are you working on something specific, or maybe my idea of cortex_m3 capabilities soft_reset would do? How about something like this: http://repo.or.cz/w/openocd/ntfreak.git/shortlog/refs/heads/stm32reset No objections to this then i will commit later today. Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
--- On Fri, 12/3/10, Spencer Oliver s...@spen-soft.co.uk wrote: On 02/12/2010 12:42, freddie_cho...@op.pl wrote: Spencer Olivers...@spen-soft.co.uk napisał(a): As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - I've lost track of the discussion here, but this sounds wrong. The only time commands should care about std (standard)/default settings is when restoring them. Otherwise, they should work with *current* settings. One bug that old versions of rset_config had was that they'd mangle unrelated settings ... e.g. you couldn't update SRST handling without trashing TRST handling (or similar; now fixed.). The two (?) reset implementation options on M3 parts should respect the same policy sanity: don't muck with options that could/should be orthogonal, and don't embed knowledge that may already be wrong, and will likely become so before long as new boards, silicon, and JTAG adapters roll out. - Dave This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. And this is an example o how I've lost the thread here. Not all Stellaris boards have SRST, since that JTAG command works on those M3 parts. Some have it, but not TRST ... So again, don't be hard-wiring incorrect facts anywhere. ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 2010-12-03 12:49, Spencer Oliver wrote: On 02/12/2010 13:15, Spencer Oliver wrote: How about something like this: http://repo.or.cz/w/openocd/ntfreak.git/shortlog/refs/heads/stm32reset No objections to this then i will commit later today. If I understand your path correctly - it removes option srst and software reset options never override hardware resets if those are available? 4\/3!! ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 03/12/2010 19:13, Freddie Chopin wrote: On 2010-12-03 12:49, Spencer Oliver wrote: On 02/12/2010 13:15, Spencer Oliver wrote: How about something like this: http://repo.or.cz/w/openocd/ntfreak.git/shortlog/refs/heads/stm32reset No objections to this then i will commit later today. If I understand your path correctly - it removes option srst and software reset options never override hardware resets if those are available? Correct. Cheers Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 03/12/2010 17:24, David Brownell wrote: --- On Fri, 12/3/10, Spencer Olivers...@spen-soft.co.uk wrote: On 02/12/2010 12:42, freddie_cho...@op.pl wrote: Spencer Olivers...@spen-soft.co.uk napisał(a): As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - I've lost track of the discussion here, but this sounds wrong. The only time commands should care about std (standard)/default settings is when restoring them. Otherwise, they should work with *current* settings. One bug that old versions of rset_config had was that they'd mangle unrelated settings ... e.g. you couldn't update SRST handling without trashing TRST handling (or similar; now fixed.). The two (?) reset implementation options on M3 parts should respect the same policy sanity: don't muck with options that could/should be orthogonal, and don't embed knowledge that may already be wrong, and will likely become so before long as new boards, silicon, and JTAG adapters roll out. Sorry i am slightly confused this patch as such has nothing todo with the standard reset_config. - Dave This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. And this is an example o how I've lost the thread here. Not all Stellaris boards have SRST, since that JTAG command works on those M3 parts. Some have it, but not TRST ... So again, don't be hard-wiring incorrect facts anywhere. Again you need to explain - i am totally confused by your whole reply. Cheers Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 2010-12-03 23:29, Spencer Oliver wrote: On 03/12/2010 19:13, Freddie Chopin wrote: On 2010-12-03 12:49, Spencer Oliver wrote: On 02/12/2010 13:15, Spencer Oliver wrote: How about something like this: http://repo.or.cz/w/openocd/ntfreak.git/shortlog/refs/heads/stm32reset No objections to this then i will commit later today. If I understand your path correctly - it removes option srst and software reset options never override hardware resets if those are available? Correct. Then the idea is pretty fine (; I believe that you've coded that right ; Seems OK for me. My oroginal patch (adding soft reset configuration for stm32.cfg) or the version from your repository will still be useful and I hope that you'll commit that too. 4\/3!! ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
--- On Fri, 12/3/10, Spencer Oliver s...@spen-soft.co.uk wrote: Date: Friday, December 3, 2010, 2:34 PM On 03/12/2010 17:24, David Brownell wrote: --- On Fri, 12/3/10, Spencer Olivers...@spen-soft.co.uk wrote: Sorry i am slightly confused this patch as such has nothing to do with the standard reset_config. Good -- matches my original impression from back a while. I think Freddie's comment (that it does have something to do with it) made confusion, then. - Dave This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. And this is an example o how I've lost the thread here. Not all Stellaris boards have SRST, since that JTAG command works on those M3 parts. Some have it, but not TRST ... it == SRST, used instead of that JTAG op (register write ISTR) that does a similar reset. (Though some might put it the other way around: the reset-via-JTAG compensating lack-of-SRST. And to be clear, that reset-via-JTAG is indeed an implementation option, which not all M3 chips handle like Stellaris ones do (which is IMO the preferable option.) Until perhaps recently, OpenOCD preferred that JTAG op for Stellaris chips (based on variant or whatever, I forget). ISTR the other reset-via-JTAG was cpu-only and less useful for the rest of the chip or system. So again, don't be hard-wiring incorrect facts anywhere. Again you need to explain - A number of emails in this thread advocated some stuff that seemed plain wrong to me, like the above romoval of SRST config even in cases where that would be wrong. i am totally confused by your whole reply. That would be explained by my trusting Freddie's comment to be accurate, i while it was instead ncorrect at the levels I was commenting on. Just keep things orthogonal; I originally thought you were, ergo my confusion when I saw several posts quoting otherwise... Bad premises lead to much confusion, as I'm sure you know. - Dave ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config = ...quot; to stm32.cfg
Spencer Oliver s...@spen-soft.co.uk napisał(a): I thought we has decided to work on the ability to make openocd a bit more reset aware for cortex-m3. This is just for now, as maybe a new bug-fix release will be made. This fixes the case for some users, but will break it for others. For whom does that break anything? STM32 supports sysresetreq, so why not use it in target config file? For the moment i would say hold off, and lets come up with a better solution that makes everyone happy. This is a temporary fix that breaks nothing until something else will be introduced. 4\/3!! ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config = ...quot; to stm32.cfg
On 02/12/2010 11:55, freddie_cho...@op.pl wrote: Spencer Olivers...@spen-soft.co.uk napisał(a): I thought we has decided to work on the ability to make openocd a bit more reset aware for cortex-m3. This is just for now, as maybe a new bug-fix release will be made. This fixes the case for some users, but will break it for others. For whom does that break anything? STM32 supports sysresetreq, so why not use it in target config file? As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - this has undesired effects for people who expect srst for example to work. We need to change that functionality then adding your patch would be ok. This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. I am looking at this at the moment, as i would like a solution. Cheers Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
Spencer Oliver s...@spen-soft.co.uk napisał(a): As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - this has undesired effects for people who expect srst for example to work. We need to change that functionality then adding your patch would be ok. This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. I am looking at this at the moment, as i would like a solution. Are you working on something specific, or maybe my idea of cortex_m3 capabilities soft_reset would do? 4\/3!! ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config= = ...quot; to stm32.cfg
On 02/12/2010 12:42, freddie_cho...@op.pl wrote: Spencer Olivers...@spen-soft.co.uk napisał(a): As we know the current behaviour of cortex_m3 reset_config is to override the std 'reset_config' setting - this has undesired effects for people who expect srst for example to work. We need to change that functionality then adding your patch would be ok. This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. I am looking at this at the moment, as i would like a solution. Are you working on something specific, or maybe my idea of cortex_m3 capabilities soft_reset would do? How about something like this: http://repo.or.cz/w/openocd/ntfreak.git/shortlog/refs/heads/stm32reset Cheers Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] [PATCH 1/2] add quot; cortex m3 reset config = ...quot; to stm32.cfg
Spencer Oliver wrote: This does create an issue for all the luminary scripts however. They would need the 'reset_config srst_only' removing and this would break all the older scripts in the wild. With a significant improvement to our data model I think it's OK to break existing scripts. (They will still work the same with the old version of the code, of course.) //Peter ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development