aboutsummaryrefslogtreecommitdiff
path: root/README.Windows
diff options
context:
space:
mode:
authorTim Newsome <tim@sifive.com>2021-03-05 15:52:33 -0800
committerGitHub <noreply@github.com>2021-03-05 15:52:33 -0800
commit50a5971be27ff11061776afc8014bbff22870b7c (patch)
treed8fe6a9d879346ba645d8eb54de048117e3f8344 /README.Windows
parentd57ab0b6321a024c3a83e0acb33eeab3b461b607 (diff)
downloadriscv-openocd-50a5971be27ff11061776afc8014bbff22870b7c.zip
riscv-openocd-50a5971be27ff11061776afc8014bbff22870b7c.tar.gz
riscv-openocd-50a5971be27ff11061776afc8014bbff22870b7c.tar.bz2
RISC-V Freertos support (#582)
Support reading names/status of all threads, as well as all registers that are stored on the stack. Limited to RV32, no FPU. --- * WIP Change-Id: I09417c2e45748504be449d74c39ae0b6b311e277 * WIP Change-Id: I975fa2cabbf43ccf64f5162337c394f9c8e3017f * Import rbtreehash-list from gnulib. The main change to get this to build was to remove 3 includes from config.h (actual code change in configure.ac) because lib/Makefile.am doesn't contain the correct flags to find the files referenced there. Instead I sprinkled necessary includes throughout the source code. This feels like less of a hack regardless, so hopefully that's OK. I'm not actually using the new library. Just got it to build. Change-Id: I824000d8be0b6f58b6f2036498b37c33f453515a * Actually use linkedhash_map. Moved some files around to get it to link. Also note I'm using a different module than before. This is the one I want (I think right now). Change-Id: I6161bffd4b5f916602c33c1930be6e061cefe982 * Properly track TCB/threadid mappings. Change-Id: I725abb96f880745d78c5634d5faff7385c2773e1 * OpenOCD no longer crashes reading rv32 freertos regs Change-Id: Ia84502dbf007145995d4fba8661153ab7f58f26f * WIP The register values reported for threads that aren't the current thread look believable to me. Change-Id: I94b109565c8cc2029fa77657a7fc10291bcb36e3 * Correctly mark the current thread. Change-Id: Id94ababb55a222292090e6465e47ebf92ca26291 * Try to make the build pass. Change-Id: I0fddd10fe22c013464f9a1e106cd21470fa7afe1 Signed-off-by: Tim Newsome <tim@sifive.com> * Exclude new gnulib files. Change-Id: I8b95615908034124f2236422771b5079f3304e37 Signed-off-by: Tim Newsome <tim@sifive.com> * Style fixes. Change-Id: I4aef0b1d0b0e366893c740ab89756fe8ea033ddb Signed-off-by: Tim Newsome <tim@sifive.com> * Don't include string.h. It breaks the i686-w64-mingw32-gcc build, which complains: error: incompatible implicit declaration of built-in function ‘strndup’ Change-Id: I8d758fe092efa503e015f71f34721f2c44632516 Signed-off-by: Tim Newsome <tim@sifive.com> * Hopefully fix mingw32 build. Change-Id: I8703b834b5679588b3aa6602ae4add7258dbd879 Signed-off-by: Tim Newsome <tim@sifive.com> * Include winsock2 in replacements. Change-Id: I77cfc90736c771a3cdefb39062e6c5b59de52cd5 Signed-off-by: Tim Newsome <tim@sifive.com> * Zero now gets the correct value. Change-Id: Ia7da043439a82081629b8a5991ed8cbc382d5ac8 Signed-off-by: Tim Newsome <tim@sifive.com> * Accommodate non-general regs on the stack. Also refactor FreeRTOS a little to separate out target-specific code from target-indepent code. Change-Id: Icc74d85b24f35d069be091e32e23144573560e9f * All registers now read sane values. It appears that FreeRTOS wastes a space on the stack, where x0 would be saved. Am I missing something? Correctly read mstatus as it is saved on the stack as well. This same mechanism should also work for FPU registers, although there's more work to be done before we get there. Change-Id: Iabacc3af2ab368aa7b9090c1ff719451a087b5ed Signed-off-by: Tim Newsome <tim@sifive.com>
Diffstat (limited to 'README.Windows')
0 files changed, 0 insertions, 0 deletions