6db8afb9d6
Flush the current cache first
2018-05-30 21:31:53 -07:00
9d975bf8f0
Be a bit stricter about fences
2018-05-30 21:30:51 -07:00
d6a0023bab
Revert "Don't show a pre-attached connection in GDB"
...
This reverts commit de71abeaba
2018-05-30 21:01:08 -07:00
de71abeaba
Don't show a pre-attached connection in GDB
2018-05-30 20:55:27 -07:00
2c30659de8
Define a debug reason from the start
2018-05-30 19:40:47 -07:00
07fcaa226b
Keeep the RTOS hartid and the current hart in sync
2018-05-30 18:19:29 -07:00
0bf2976aa3
A vCont;c packet applies to all harts?
2018-05-30 17:56:45 -07:00
91e544d66c
Invalidate the register cache more often
2018-05-30 17:48:48 -07:00
d2cd725dd3
Invalidate the register cache when touching another hart
...
The 0.13 code now caches registers interally, so when reading registers
on a diferent hart we need to invalidate the cache.
2018-05-30 17:36:22 -07:00
6d9e69499f
Don't rely on the RTOS hartid for the register cache
2018-05-30 07:04:14 -07:00
06bc6cccd4
Revert "Don't make callbacks from riscv_openocd_{halt,resume}"
...
This reverts commit bfddd9af64
2018-05-30 06:26:57 -07:00
bfddd9af64
Don't make callbacks from riscv_openocd_{halt,resume}
...
I'm not sure why this would be an issue, but it looks like for some
reason this is causing extraneous halt messages in Eclipse.
2018-05-30 06:20:16 -07:00
a8f45d36a8
I guess just set all harts, not any hart
2018-05-30 06:07:56 -07:00
395592ece3
Revert "Try saving the halted state when examining a target"
...
This reverts commit f1e8dee522
2018-05-30 06:00:30 -07:00
2ec501a8b3
Revert "double result"
...
This reverts commit 1b227f1f49
2018-05-30 06:00:14 -07:00
1b227f1f49
double result
2018-05-30 05:56:42 -07:00
f1e8dee522
Try saving the halted state when examining a target
2018-05-30 05:54:56 -07:00
9b19e10f8d
Don't set RTOS hartid 0
2018-05-30 05:02:05 -07:00
3d8393c844
trnprintf
2018-05-30 04:55:24 -07:00
6d2f17a823
Handle the qC packet
2018-05-30 04:53:54 -07:00
900af90662
sntrnmp
2018-05-30 04:45:39 -07:00
f6d6ed85ab
whoops
2018-05-30 04:44:14 -07:00
9889ccb1bb
Pick an arbitrary hart ID for the target
2018-05-30 04:30:57 -07:00
045d5b0142
Handle qTStatus
2018-05-30 04:30:48 -07:00
330b11309d
debug GDB IO
2018-05-30 04:24:10 -07:00
65e5116229
Revert "Print allt he GDB messages back"
...
This reverts commit 4fc0e7ac0d
2018-05-30 04:23:49 -07:00
7d693c6983
Revert "iwhoops"
...
This reverts commit d3c9ed508d
2018-05-30 04:23:47 -07:00
d3c9ed508d
iwhoops
2018-05-30 02:56:22 -07:00
4fc0e7ac0d
Print allt he GDB messages back
2018-05-30 02:53:52 -07:00
ec42c4300e
Whoops
2018-05-30 02:25:30 -07:00
47731c68d2
If we don't know which thread should be halted then just don't set one
2018-05-30 02:22:59 -07:00
ab5dbc6168
Print signal replies
2018-05-30 02:17:05 -07:00
a7e00a8e72
More debugging
2018-05-30 01:42:20 -07:00
fccc20ad7a
More debug info
2018-05-30 01:29:15 -07:00
3c00bd8ff2
Enable debug during the poll
2018-05-30 01:11:17 -07:00
0be30cc58c
Print a bit more when a hart is halted
...
I think I'm getting some "wrong hart halted" issues here.
2018-05-30 01:03:02 -07:00
755bf8d558
Revert "Don't raise HALT when we're examining a target"
...
This reverts commit dd382bb6fb
2018-05-30 00:59:41 -07:00
a2d118f8e4
Revert "Don't make callbacks from riscv_openocd_{halt,resume}"
...
This reverts commit fb54cc4fa5
2018-05-30 00:33:19 -07:00
dd382bb6fb
Don't raise HALT when we're examining a target
2018-05-30 00:29:36 -07:00
fb54cc4fa5
Don't make callbacks from riscv_openocd_{halt,resume}
...
I'm not sure why this would be an issue, but it looks like for some
reason this is causing extraneous halt messages in Eclipse.
2018-05-30 00:20:45 -07:00
e373719ac1
Don't halt whenever GDB attaches
2018-05-30 00:05:27 -07:00
3ce353cafa
Revert "Don't make callbacks from riscv_openocd_{halt,resume}"
...
This reverts commit 43092445df
2018-05-30 00:05:16 -07:00
43092445df
Don't make callbacks from riscv_openocd_{halt,resume}
...
I'm not sure why this would be an issue, but it looks like for some
reason this is causing extraneous halt messages in Eclipse.
2018-05-29 23:34:56 -07:00
6b85d0945c
Revert "Don't raise HALT when we're examining a target"
...
This reverts commit b39d196489
2018-05-29 23:27:41 -07:00
b39d196489
Don't raise HALT when we're examining a target
2018-05-29 23:18:23 -07:00
d7bb150086
Move JTAG batch printing to DEBUG_LVL_IO
2018-05-29 19:59:14 -07:00
052b4e3142
Don't print verbose messages when polling
...
I'm not 100% sure what the right thing to do here is, but I've found
that when I'm not debugging a polling issue there's way too much
verbosity in the debug level as it currently stands.
2018-05-29 19:51:00 -07:00
ab7ab8a867
Merge pull request #261 from riscv/trigger_enum
...
Delay trigger enumeration until it's required.
2018-05-25 11:52:10 -07:00
c3ffbc66e6
Merge pull request #257 from riscv/comment
...
Comment riscv_set_register, register_write_direct
2018-05-22 14:39:28 -07:00
b629bbeade
Delay trigger enumeration until it's required.
...
This improves startup time, which is important when connecting to
simulators. One problem is that triggers that are set when the debugger
connects are not cleared until enumeration happens. Execution may halt
due to a trigger set by a previous debug session, which could confuse
the user. If this happens, triggers will be instantly enumerated, so it
will only happen once per session.
Change-Id: I3396f713f16980a8b74745a1672fe8b8a2d4abae
2018-05-22 13:07:25 -07:00
Palmer Dabbelt
Tim Newsome