Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Jonathan Dumaresq
Hi Spen, This is looking good. Now It's break when the ars is read or write. Thanx for you time and help ! Jonathan Spencer Oliver wrote: > On 25 January 2013 15:25, Spencer Oliver wrote: >> >> For some reason the DWT logic was disabled - the icdi must have been >> enabling it itself. >> >>

Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Jonathan Dumaresq
Hi Spen, I'll give it a try ! Jonathan Spencer Oliver wrote: > On 25 January 2013 15:25, Spencer Oliver wrote: >> >> For some reason the DWT logic was disabled - the icdi must have been >> enabling it itself. >> >> So looking into this further it appears to be an OpenOCD issue. I >> will post

Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Spencer Oliver
On 25 January 2013 15:25, Spencer Oliver wrote: > > For some reason the DWT logic was disabled - the icdi must have been > enabling it itself. > > So looking into this further it appears to be an OpenOCD issue. I will > post a patch to gerrit that you can test. > http://openocd.zylin.com/1122 Sp

Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Spencer Oliver
On 25 January 2013 14:48, Jonathan Dumaresq wrote: >> >> OK i have looked into it and it seems the stlink is overwriting the >> watchpoint config that we are writing to the target. >> I tested the ti-icdi as it uses the same driver as the stlink and >> that works as expected. > > At least we have

Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Jonathan Dumaresq
Hi Spen, Spencer Oliver wrote: > On 23 January 2013 12:14, Spencer Oliver wrote: >> On 22 January 2013 18:14, Jonathan Dumaresq >> wrote: >>> Hi Spen, >>> >>> I did this as a test >>> >>> volatile uint32_t *test = (uint32_t *)0x20001000; >>> >>> while(1) >>> *

Re: [OpenOCD-devel] Watchpoint

2013-01-25 Thread Spencer Oliver
On 23 January 2013 12:14, Spencer Oliver wrote: > On 22 January 2013 18:14, Jonathan Dumaresq wrote: >> Hi Spen, >> >> I did this as a test >> >> volatile uint32_t *test = (uint32_t *)0x20001000; >> >> while(1) >> *test = *test + 1; >> >> Whould I been able to brea

Re: [OpenOCD-devel] Watchpoint

2013-01-23 Thread Spencer Oliver
On 22 January 2013 18:14, Jonathan Dumaresq wrote: > Hi Spen, > > I did this as a test > > volatile uint32_t *test = (uint32_t *)0x20001000; > > while(1) > *test = *test + 1; > > Whould I been able to break on access to memory location 0x20001000 ? > > wp 0x20001000

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Jonathan Dumaresq
Hi Spen, I did this as a test volatile uint32_t *test = (uint32_t *)0x20001000; while(1) *test = *test + 1; Whould I been able to break on access to memory location 0x20001000 ? wp 0x20001000 4 rw Nothing stop ! Jonathan

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Jonathan Dumaresq
Hi Spen, > > as you are writing 32bit words above the you should really use wp > 0x2000 4 rw > Is it possible to have watch point that will trigger at specific address without specifiy the access of it ? (1 2 or 4 bytes wide ? When you have tested it is it with gdb ? Or directly by telnet

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Spencer Oliver
On 22 January 2013 16:06, Jonathan Dumaresq wrote: > Hi Spen, > > Finnaly I got it checked out with this > > git fetch http://openocd.zylin.com/openocd refs/changes/31/931/3 && git > checkout FETCH_HEAD > > But it's not seem to work ethier. > > Here my tes on the stm32F4 target > > mdw 0x2000

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Jonathan Dumaresq
Hi Spen, Finnaly I got it checked out with this git fetch http://openocd.zylin.com/openocd refs/changes/31/931/3 && git checkout FETCH_HEAD But it's not seem to work ethier. Here my tes on the stm32F4 target mdw 0x2000 1 0x2000: 00600321 mww 0x2000 0 1 mdw 0x2000 1 0x2000

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Spencer Oliver
On 22 January 2013 12:55, Jonathan Dumaresq wrote: >> >> Working when i fixed a regression below: >> http://openocd.zylin.com/931 >> >> I presume you are using latest master ? >> >> Spen > > This doesn't seem to work better with the latest Git head. > Try checking out OpenOCD when the patch was m

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Jonathan Dumaresq
Hi spen, Spencer Oliver wrote: > On 21 January 2013 20:11, Jonathan Dumaresq > wrote: >> Hi, >> >> I try to use watchpoint to find a problem on my target. >> >> Stm32f4 device. >> >> Before I start the application, I execute the commands in a telnet >> windows >> >> wp 0x20005414 4 a >> >>

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Jonathan Dumaresq
Hi Spen, Spencer Oliver wrote: > On 21 January 2013 20:11, Jonathan Dumaresq > wrote: >> Hi, >> >> I try to use watchpoint to find a problem on my target. >> >> Stm32f4 device. >> >> Before I start the application, I execute the commands in a telnet >> windows >> >> wp 0x20005414 4 a >> >>

Re: [OpenOCD-devel] Watchpoint

2013-01-22 Thread Spencer Oliver
On 21 January 2013 20:11, Jonathan Dumaresq wrote: > Hi, > > I try to use watchpoint to find a problem on my target. > > Stm32f4 device. > > Before I start the application, I execute the commands in a telnet windows > > wp 0x20005414 4 a > > When I look into the wp list I see this > > address: 0x2

[OpenOCD-devel] Watchpoint

2013-01-21 Thread Jonathan Dumaresq
Hi, I try to use watchpoint to find a problem on my target. Stm32f4 device. Before I start the application, I execute the commands in a telnet windows wp 0x20005414 4 a When I look into the wp list I see this address: 0x20005614, len: 0x0004, r/w/a: 2, value: 0x, mask: 0xf F