duane>> [[ started this: about creating a memory range to determine
breakpoints, because CortexM macro is wrong headed ]]
Freddie>> [[ points out using the flash memory area … ]]
duane>> [ That means the target has to be probed - cannot always probe ]
duane>> [ Not all targets have a “flash driver” so they cannot be probed ]
Andreas>> [ Points out that most ‘workspaces’ cover all RAM, and this is a
configuration option ]
duane>> [ points out there are many types of memory maps - that are not
probable, not just work areas]
So - my new proposal is this:
What if we create a command:
$_TARGET memmap SHOW
print a nice table
$_TARGET memmap CLEAR
set entire range as NOTPRESENT
$_TARGET memmap INSERT TYPE START LENGTH
Insert a new range with specific type
Where TYPE =
RAM = generic ram
Should this support suffixes?
RAM.8 = ram accessible in 8bit mode only
RAM.16 - 16 bit mode only
RAM.32
ROM < and suffixes >
FLASH <and suffixes >
HWREGS <and suffixes>
NOTPRESENT
The target defaults will be:
ALL of memory = RAM
(when the target structure is allocated and initialized)
For Cortex M, there will be a default range
My changes will support this.
Done in the arch init code
Other targets [I’m not going to do this, others can]
Option 1: During their arch init - can modify the map
Option 2: After flash probe occurs can modify the map
Flash drivers - once probing is complete they can choose to call the
mmap functions below.
Not sure - should the “work areas” modify the target memmap?
I don’t think so - because by default those areas would by default
already be marked as RAM
new types:
enum MEM_RANGE_TYPE /* per above */
struct target_mem_range /* details below */
/* this is a longer term issue that needs to be started */
typedef uint64_t target_addres_t; /* think 64bit arch!!!! */
new functions with this signature:
BKPT_TYPE target_breakpoint_by_address( target, address )
(replacing the cortex_m macro that caused this mess)
BKPT_TYPE = ram - for any ram type thing
BLPT_HARD = all otters
New functions:
/* set all to not present */
target_mmap_clear( target );
/* split current map and insert this new map */
target_mmap_insert( target, TYPE, START, LENGTH )
implementation will be an array of:
/* pointer to array of this in the target structure */
struct target_mmap_range {
enum RAMTYPE type;
target_addr_t start;
target_off_t len;
};
/* and this in the target structure */
int n_target_mmaps;
Thoughts??
------------------------------------------------------------------------------
Is your legacy SCM system holding you back? Join Perforce May 7 to find out:
• 3 signs your SCM is hindering your productivity
• Requirements for releasing software faster
• Expert tips and advice for migrating your SCM now
http://p.sf.net/sfu/perforce
_______________________________________________
OpenOCD-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openocd-devel
