dlbeer
unable to step device after reset
I'm unable to do more than one 'step' command in mspdebug after issuing a reset.
Bug reproduced on version 0.22 and 0.23 as of GIT 38ea6143cdf9e63a8f200af25b2d4ec39d62d066
Evidence from console:
mspdebug rf2500 MSPDebug version 0.22 - debugging tool for MSP430 MCUs Copyright (C) 2009-2013 Daniel Beer [email protected] This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Trying to open interface 1 on 072
rf2500: warning: can't detach kernel driver: No data available
Initializing FET...
FET protocol version is 30001000
Set Vcc: 3000 mV
Configured for Spy-Bi-Wire
fet: FET returned error code 4 (Could not find device or device not supported)
fet: command C_IDENT1 failed
Using Olimex identification procedure
Device ID: 0x2452
Code start address: 0xe000
Code size : 8192 byte = 8 kb
RAM start address: 0x200
RAM end address: 0x2ff
RAM size : 256 byte = 0 kb
Device: MSP430G2xx2
Number of breakpoints: 2
fet: FET returned NAK
warning: device does not support power profiling
Chip ID data: 24 52
(mspdebug) reset
(mspdebug) step
( PC: 0f806) ( R4: 04fff) ( R8: 0fade) (R12: 00079)
( SP: 00280) ( R5: 00000) ( R9: 07f7f) (R13: 09fff)
( SR: 00000) ( R6: 00000) (R10: 0efff) (R14: 0bded)
( R3: 00000) ( R7: 0ffa3) (R11: 0ff4b) (R15: 0fe03)
0xf806:
0f806: 31 40 80 02 MOV #0x0280, SP
0f80a: b0 12 94 f8 CALL #0xf894
0f80e: b0 12 a2 f8 CALL #0xf8a2
0f812: f2 42 29 00 MOV.B #0x0008, &0x0029
(mspdebug) step
fet: FET returned error code 16 (Could not single step device)
fet: failed to restart CPU
On Sun, Nov 22, 2015 at 08:51:25PM -0800, Adam Di Carlo wrote:
I'm unable to do more than one 'step' command in mspdebug after issuing a reset.
Bug reproduced on version 0.22 and 0.23 as of GIT 38ea6143cdf9e63a8f200af25b2d4ec39d62d066
Evidence from console:
mspdebug rf2500 MSPDebug version 0.22 - debugging tool for MSP430 MCUs Copyright (C) 2009-2013 Daniel Beer [email protected] This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Trying to open interface 1 on 072 rf2500: warning: can't detach kernel driver: No data available Initializing FET... FET protocol version is 30001000 Set Vcc: 3000 mV Configured for Spy-Bi-Wire fet: FET returned error code 4 (Could not find device or device not supported) fet: command C_IDENT1 failed Using Olimex identification procedure Device ID: 0x2452 Code start address: 0xe000 Code size : 8192 byte = 8 kb RAM start address: 0x200 RAM end address: 0x2ff RAM size : 256 byte = 0 kb Device: MSP430G2xx2 Number of breakpoints: 2 fet: FET returned NAK warning: device does not support power profiling Chip ID data: 24 52 (mspdebug) reset
Hi Adam,
Do you know whether this bug is present in older versions? Does single-stepping work if you don't issue a reset?
Also, if you happen to have the library handy, would you mind trying with the tilib driver?
Cheers, Daniel
Daniel Beer [email protected] http://dlbeer.co.nz/ PGP: BA6E 0B26 1F89 246C E3F3 C910 1E58 C43A 160A 553B
Do you know whether this bug is present in older versions?
I'm pretty sure I remember it from 0.21 as well. I tried to rollback to git 5477cd78d87bb061f54766efdceb988ddcdd6ea5 and compile, but I'm getting a compilation error on util/sport.c: util/sport.c: In function ‘sport_read’: util/sport.c:162:10: error: variable ‘tv’ has initializer but incomplete type struct timeval tv = {
amongst others....
I'll see about how to use the TI libraries...
Does single-stepping work if you don't issue a reset?
It does indeed!
Sadly I'm unable to attempt this with tilib:
mspdebug tilib MSPDebug version 0.22 - debugging tool for MSP430 MCUs Copyright (C) 2009-2013 Daniel Beer [email protected] This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
MSP430_GetNumberOfUsbIfs No unused FET found.
I did manage to reproduce the problem in mspdebug 0.19, using 'mspdebug rf2500'. It shows the problem just as version 0.22 and the latest do -- stepping works fine, but after reset, you can only step once.
On Mon, Nov 23, 2015 at 07:16:05AM -0800, Adam Di Carlo wrote:
Do you know whether this bug is present in older versions?
I'm pretty sure I remember it from 0.21 as well. I tried to rollback to git 5477cd78d87bb061f54766efdceb988ddcdd6ea5 and compile, but I'm getting a compilation error on util/sport.c: util/sport.c: In function ‘sport_read’: util/sport.c:162:10: error: variable ‘tv’ has initializer but incomplete type struct timeval tv = {
amongst others....
I'll see about how to use the TI libraries...
Does single-stepping work if you don't issue a reset?
It does indeed!
Hmm... I don't have your chip here, but if you want to try something, I'd suggest messing with the reset type. In fet_ctl(), in drivers/fet_core.c, there is a switch statement. The case for DEVICE_CTL_RESET is where the reset is issued.
There is a parameter, curently set to FET_RESET_ALL, which controls how the reset is done. These are defined near the top of the file. What happens if you change it to FET_RESET_RST (or something else)?
Cheers, Daniel
Daniel Beer [email protected] http://dlbeer.co.nz/ PGP: BA6E 0B26 1F89 246C E3F3 C910 1E58 C43A 160A 553B
Cool, I'll try that out! Busy right now but I'll frob around at this and see what I can do...
Tried FET_RESET_RST -- no go.
Ditto FET_RESET_PUC, FET_RESET_VCC -- no joy.
I want to make clear that reset is not totally broken. For instance, I can do:
reset run
and all works fine. However, if I Ctrl-C and then try to step, I'll get error 16. So, only the step which is run after another command fails.
Let me know what else we can try. Also, I'm available on IRC, I'm 'adicarlo' on freenode.
Do you have a launchpad to work with? If not, I could buy you one and send it to you direct! They're cheap....
On Tue, Nov 24, 2015 at 03:49:18PM -0800, Adam Di Carlo wrote:
Tried FET_RESET_RST -- no go.
Ditto FET_RESET_PUC, FET_RESET_VCC -- no joy.
I want to make clear that reset is not totally broken. For instance, I can do:
reset run
and all works fine. However, if I Ctrl-C and then try to step, I'll get error 16. So, only the step which is run after another command fails.
Let me know what else we can try. Also, I'm available on IRC, I'm 'adicarlo' on freenode.
Do you have a launchpad to work with? If not, I could buy you one and send it to you direct! They're cheap....
Hi Adam,
That's quite odd. Does it make any different if you try "regs" between "reset" and "step"?
I do have a launchpad here, but my main constraint is time, unfortunately. I'll do what I can to help, but I might not be able to investigate too much directly.
Cheers, Daniel
Daniel Beer [email protected] http://dlbeer.co.nz/ PGP: BA6E 0B26 1F89 246C E3F3 C910 1E58 C43A 160A 553B
That's quite odd. Does it make any different if you try "regs" between "reset" and "step"?
It does! So my original recipe, still happening, is: reset step step -- bewm
But if I insert 'regs', the first step will blow up: reset regs step -- bewm
Another clue which might be interesting: reset step step -- bewm reset step -- works!! step -- bewm
Always the 2nd operation after the reset! Although some operations work over and over: reset regs regs Well, regs always works.
Also, this works: reset run
Also this works, although I expected it to fail: reset step run
Note that all this behavior is 100% consistent.
On Wed, Nov 25, 2015 at 09:33:05AM -0800, Adam Di Carlo wrote:
That's quite odd. Does it make any different if you try "regs" between "reset" and "step"?
It does! So my original recipe, still happening, is: reset step step -- bewm
But if I insert 'regs', the first step will blow up: reset regs step -- bewm
Another clue which might be interesting: reset step step -- bewm reset step -- works!! step -- bewm
Always the 2nd operation after the reset! Although some operations work over and over: reset regs regs Well, regs always works.
Also, this works: reset run
Also this works, although I expected it to fail: reset step run
Note that all this behavior is 100% consistent.
Hmm. Ok, just a semi-wild guess, but what if you modify cmd_step() so that it calls device_ctl(DEVICE_CTL_HALT) after doing the single-step (DEVICE_CTL_STEP)?
Cheers, Daniel
Daniel Beer [email protected] http://dlbeer.co.nz/ PGP: BA6E 0B26 1F89 246C E3F3 C910 1E58 C43A 160A 553B
Well, I'm just cargo culting here, not really sure if this is what you asked for, but this is what I tried:
diff --git a/drivers/fet_core.c b/drivers/fet_core.c index d89f607..9016da1 100644 --- a/drivers/fet_core.c +++ b/drivers/fet_core.c @@ -703,6 +703,12 @@ int fet_ctl(device_t dev_base, device_ctl_t action) if (do_run(dev, FET_RUN_STEP) < 0) return -1;
-
power_end(dev); -
if (fet_proto_xfer(&dev->proto, C_STATE, NULL, 0, 1, 1) < 0) { -
printc_err("fet: failed to halt CPU\n"); -
return -1; -
}- for (;;) { device_status_t status = fet_poll(dev_base);
But sadly, no change in behavior can be detected.
On Wed, Nov 25, 2015 at 08:48:25PM -0800, Adam Di Carlo wrote:
Well, I'm just cargo culting here, not really sure if this is what you asked for, but this is what I tried:
diff --git a/drivers/fet_core.c b/drivers/fet_core.c index d89f607..9016da1 100644 --- a/drivers/fet_core.c +++ b/drivers/fet_core.c @@ -703,6 +703,12 @@ int fet_ctl(device_t dev_base, device_ctl_t action) if (do_run(dev, FET_RUN_STEP) < 0) return -1;
power_end(dev);if (fet_proto_xfer(&dev->proto, C_STATE, NULL, 0, 1, 1) < 0) {printc_err("fet: failed to halt CPU\n");return -1;}
- for (;;) { device_status_t status = fet_poll(dev_base);
But sadly, no change in behavior can be detected.
What if you put it after the poll loop (for (;;) { ... })?
Failing that, I'm a bit stumped. I have no documentation for the FET protocol except what I've tried and got working.
Daniel Beer [email protected] http://dlbeer.co.nz/ PGP: BA6E 0B26 1F89 246C E3F3 C910 1E58 C43A 160A 553B
What if you put it after the poll loop (for (;;) { ... })?
Aye, tried that, no luck.
Failing that, I'm a bit stumped. I have no documentation for the FET protocol except what I've tried and got working.
Well, I'm pretty new to this sort of debugging, but I'm willing to try to take a look. Is possible to get me started with a citation of the relevant document/page ?