* Deal with vlenb being unreadable.
Instead of exiting during examine(), spit out a warning, and don't
expose the vector data registers. We do provide access to the vector
CSRs, because maybe they do work? It's just that we have no idea what
the size of the data registers is.
Change-Id: I6e9ffeb242e2e22fc62cb1b50782c2efb4ace0bd
* WIP
Change-Id: I46292eefe537aeaf72bdd44e4aa58298b5120b00
* Use the correct thread for memory accesses.
Previously, OpenOCD would perform RTOS memory accesses through the first
thread in the RTOS. This doesn't work if different threads have a
different memory view. For instance if `-rtos hwthread` is used, each
configured core could have address translation configured differently.
Change-Id: I61328c8f50065ecba5ce1797dbeaee482812f799
Fake step is a hack introduced to make things work with real RTOSs that
have a concept of a current thread. The hwthread rtos always has access
to all threads, so doesn't need it.
This fixes a bug when running my MulticoreRegTest against HiFive
Unleashed where OpenOCD would return the registers of the wrong thread
after gdb stepped a hart.
Change-Id: I64f538a133fb078c05a0c6b8121388b0b9d7f1b8
This patch adds "hwthread", a pseudo rtos that represents cpu cores
in an SMP system as threads to gdb. This allows to debug SMP
system kernels in a more sensible manner and removes the current
atrocities of switching gdb manually between CPU cores to update
the context.
Change-Id: Ib781c6c34097689d21d9e02011e4d74a4a742379
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3999
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Graham Sanderson <graham.sanderson@gmail.com>
* Implement riscv_get_thread_reg().
This is necessary because riscv_get_gdb_reg_list() now reads all
registers, which ended up causing `-rtos riscv` to read all registers
whenever one was requested (because the register cache is wiped every
time we switch to a different hart).
CustomRegisterTest went from 1329s to 106s.
Change-Id: I8e9918b7a532d44bca927f67aae5ac34954a8d32
* Also implement riscv_set_reg().
Now all the `-rtos riscv` tests pass again, at regular speed.
Change-Id: I55164224672d9dcc9eb4d1184b47258ff3c2cff1
* Better error messages.
Change-Id: I4125f9a54750d9d0ee22c4fa84b9dd3f5af203f5
* Add target_get_gdb_reg_list_noread().
Being explicit about what's expected gets `-rtos riscv` back to `-rtos
hwthread` time.
Change-Id: I6e57390c2fe79b5e6799bfda980d89697e2e29f7
* Revert a change I made that has no effect.
I don't understand exactly what all this test protects against, and I
shouldn't change it unless I do.
Change-Id: Ib329d4e34d65d2b38559b89b7afb3678f439ad2c
Previously the code made the assumption (which is valid for conventional
RTOSs) that special registers (e.g. CSRs) are the same across threads.
26/45 tests pass.
Change-Id: Ibb3398790d7354a995d506772375d869f608f1f0
I don't understand how it was ever possible to change the registers on a
thread that's not the current active one when a halt happened. Really
instead of the RTOS tracking what the currently selected thread is, it
would make more sense to have gdb_server do that and simply pass it
along in every call to the RTOS layer.
Now MulticoreRegTest passes.
Change-Id: I399b9b2b05a147aa6b41463714ed3a39534b1fc8
This patch adds "hwthread", a pseudo rtos that represents cpu cores
in an SMP system as threads to gdb. This allows to debug SMP
system kernels in a more sensible manner and removes the current
atrocities of switching gdb manually between CPU cores to update
the context.
Change-Id: Ib781c6c34097689d21d9e02011e4d74a4a742379
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Source: http://openocd.zylin.com/#/c/3999
Tim Newsome
Matthias Welwarsky