Hello,
When I implemented the smp, I focused on hardware breakpoints.
And later on , I noticed that the behavior was incorrect for software 
breakpoints (as you mentioned)
For software breakpoints, I was planning to  :

*         Create in target an addionnal field swbrk_target ,

*         This field will be initialized at  smp init, by selecting a dedicated 
target

In the smp list (i.e according to core id number) ,

*         And the the software breakpoint will be added/removed by using the 
swbrk_target , when smp is set to one.

Best Regards
Michel JAOUEN



From: [email protected] 
[mailto:[email protected]] On Behalf Of James Zhao
Sent: Monday, September 26, 2011 11:54 PM
To: [email protected]
Subject: [Openocd-development] software breakpoint with multicore systems

I am not sure if this is a bug or I have misunderstood something the code, but 
there seems to be a problem with software breakpoints.
When adding a software breakpoint, the function breakpoint_add() would add a 
breakpoint to all targets, if smp is on.
But since software breakpoints are done but replacing the instruction at the 
memory address with something to break when run, then why do we need to do this 
for all the cores, isn't once enough?

James

________________________________
This email message is for the sole use of the intended recipient(s) and may 
contain confidential information.  Any unauthorized review, use, disclosure or 
distribution is prohibited.  If you are not the intended recipient, please 
contact the sender by reply email and destroy all copies of the original 
message.
________________________________
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to