Update lowrisc_ibex to lowRISC/ibex@123d46b4 #20889
Conversation
msfschaffner
requested review from
Razer6,
a-will,
andreaskurth,
nasahlpa and
vogelpi
|
CHANGE AUTHORIZED: hw/vendor/lowrisc_ibex/rtl/ibex_pmp.sv |
msfschaffner
force-pushed
the
integrated_dev
branch
from
6345679 to
003684f
Compare
|
It may be worth delaying this update until lowRISC/ibex#2129 has been merged as well, so that both FI CM improvements can be included in the same vendor-update. The S&P layer change could be reverted by applying a patch during the vendoring. This would allow to keep Ibex used in Integrated aligned with upstream with the exception of the S&P layer change. |
|
Sounds good, though we may have to be mindful of the timeline. Do you expect that change to land soon? Because if not, it may be safer to go ahead and merge this for now, and do another vendor later... |
|
After a second review round for #2129, I think we should be good to merge the fix into the Ibex repo. But it is fine for me including this fix in another vendor later. |
Update code from upstream repository https://github.com/lowRISC/ibex.git to revision 123d46b4d60068502f80c117772a279db12f5af7 * [dv] Fix paths in `merge_cov.py` (Sᴜᴘᴇʀ Lᴇᴇ) * Tweak questa timescale argument (Harry Callahan) * Fixup the questa build/sim command templates in rtl_simulation.yaml (Harry Callahan) * [rtl] Fix FI vulnerability in RF (Pascal Nasahl) * [doc] Update cosim version (Pascal Nasahl) * [util] Update check_tool_requirements.py (Gary Guo) * [rtl] Avoid name collision in ibex_pmp.sv (Rupert Swarbrick) * [dv] Fix performance counter printing in simple system (Rupert Swarbrick) * Fix spelling of separator (Rupert Swarbrick) * [dv] Add an extra key to common_project_cfg.hjson (Rupert Swarbrick) * [verilator] Slight refactor in ibex_tracer to avoid BLKSEQ warning (Rupert Swarbrick) * [verilator] Waive MULTIDRIVEN warning in ibex_tracer.sv (Rupert Swarbrick) Signed-off-by: Michael Schaffner <msf@opentitan.org>
Signed-off-by: Michael Schaffner <msf@opentitan.org>
The assertion is for the same condition as in the pinmux_strap_sampling module. However, as opposed to that module, there are some modifications to the strap_en_i signal resulting in a local version of it (strap_en) and the SVA needs to be formulated for that one instead. Signed-off-by: Michael Schaffner <msf@opentitan.org>
msfschaffner
force-pushed
the
integrated_dev
branch
from
65e2794 to
374cdbb
Compare
|
@msfschaffner PR#2129 is now merged into the Ibex repository. |
|
The decision has been taken to delay this PR until after the Darjeeling-ES.M4 deadline |
|
We do not plan to carry this improvement over to |
This vendors in the newest ibex version. The important fix we'd like to absorb in Darjeeling is the
FI vulnerability in RFpatch.Note that this purposefully does not include newer primitive changes vendored from OT -> Ibex since these could interfere with the current state on Darjeeling (e.g. the S&P layer change in the memory scrambling primitive).
Update code from upstream repository
https://github.com/lowRISC/ibex.git to revision
123d46b4d60068502f80c117772a279db12f5af7
merge_cov.py(Sᴜᴘᴇʀ Lᴇᴇ)