Patch 1.0 RTL with Lint Fixes #518
Conversation
commit e08900f Author: Michael Norris <[email protected]> Date: Thu Mar 28 11:00:58 2024 -0700 rolling back packed struct removal enabling lint on integrated TRNG files fixes for integrated TRNG lint violations commit 75ecd6c Author: Michael Norris <[email protected]> Date: Wed Mar 27 12:59:24 2024 -0700 fixing lint errors from LMS changes commit 3049b98 Author: Caleb Whitehead <[email protected]> Date: Thu Mar 28 16:44:43 2024 -0700 Squashed commit of the following: commit f03b9ba Merge: bc73a52 55f5f95 Author: Michael Norris <[email protected]> Date: Wed Mar 27 12:51:20 2024 -0700 Merge remote-tracking branch 'chips/main' into user/dev/michnorris/lint_fix commit bc73a52 Author: Michael Norris <[email protected]> Date: Wed Mar 27 12:45:42 2024 -0700 removing change from other lint pr branch, not sure where it came from. commit 697a740 Author: Michael Norris <[email protected]> Date: Wed Mar 27 12:20:42 2024 -0700 removing lint blackboxes removed packed struct post processing from generated reg files. cast width of dynamic shifts to make lint happy casting some arithmetic widths, math parameter widths etc commit a16581d Author: Michael Norris <[email protected]> Date: Tue Mar 26 15:00:06 2024 -0700 initial round of lint fixes
Same as commits: 17c0ce9459583a2250c6f292e0e468fbbf7c34 2c4b61f at user/dev/michnorris/lint_fix
calebofearth
requested review from
bharatpillilli,
JohnTraverAmd,
Nitsirks,
FerralCoder and
mborder-amd
Do you want me to run Lint and RDC on this PR? |
Keep applying fw_ov_rd_fifo_overflow instead of pulsing. This commit is a manual port of the following Opentitan commit: b454878f2528b49318ff2b604cb8d59a6e59b0ea Which is associated with lowRISC/opentitan#21640. Original commit description: This commit applies the fw_ov_rd_fifo_overflow signal in case of an overflow. It used to only send a pulse through the register but now the signal is applied until the FIFO has been emptied once and is receiving contiguous data again. This commit also aligns the documentation for the FW_OV_RD_FIFO_OVERFLOW register. The register is now not clearable by software. It instead is cleared by hardware. Signed-off-by: Miguel Osorio <[email protected]>
* Resolving unnamed generate blocks Lint warnings * Running stamp_repo.sh --------- Co-authored-by: Avirup Mullick <[email protected]>
Yes, thanks, I think there are some slight changes since last time you ran on this branch. |
Is this branch different from what we have at TOT? I mean does this have the 1.1 changes as well? Or this is just 1.0 + Lint fixes? |
Only 1.0 code + lint fixes. I just pushed updates to HW_REV_ID, as well as README+Release_Notes. The branch is good to run your checks now. Thanks! |
RDC is clean on this PR |
…width for lint fixes (#524) * fixing bugs caused during lint fixes where shifts were cast as the wrong width and bits are truncated fixed by removing the shifts and explicitly taking the bits required * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/lint_bug_fix' with updated timestamp and hash after successful run * updating smoke test to sha the entire mailbox at the start * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/lint_bug_fix' with updated timestamp and hash after successful run --------- Co-authored-by: Michael Norris <[email protected]>
* patch for kv exfiltration locking api registers from being modified by uc when data is loaded from the keyvault updating smoke tests to attempt to corrupt the kv data to test the lock * updating kv smoke test to use keyvault for block register * adding multi block hmac keyvault test content * updating keyvault section of the hardware spec to explicitly call out the key locking/clearing inside the crypto function Also detailing the requirement that each iteration of a multi block operation must program the keyvault read/write operation * corrected the expected tag to match the expected output of the hmac block * adding multiblock test to l0 and nightly directed regressions * preventing commands from being issued while key is being copied to the crypto engine * changing the masking to just cover the idle case, no need to check for data present * added busy signal to crypto engines with key access multiple busy signals trigger a fatal error zeroize keyvault reads when read has an error updated ras test to include testing crypto error case * adding new port for busy signals and crypto errors to all the unit level testbenches * fixing jtag aperture to allow access to veer jtag registers only when debug is unlocked jtag path to soc ifc registers is unchanged * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * Fix for truncated bits after casting logical shifts to the incorrect width for lint fixes (#524) * fixing bugs caused during lint fixes where shifts were cast as the wrong width and bits are truncated fixed by removing the shifts and explicitly taking the bits required * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/lint_bug_fix' with updated timestamp and hash after successful run * updating smoke test to sha the entire mailbox at the start * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/lint_bug_fix' with updated timestamp and hash after successful run --------- Co-authored-by: Michael Norris <[email protected]> * [ENV] Disable wget HSTS in ROM test (#527) * Disable HSTS (https is hardcoded in makefile, no MIM attack here) * MICROSOFT AUTOMATED PIPELINE: Stamp 'cwhitehead-msft-rom-wget-hsts-disable' with updated timestamp and hash after successful run * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * updating hardware spec for crypto error fatal error fixing some typos and doc nits updating covergroups to include new crypto error fatal error bit * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * Apply suggested feedback * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * Removed multiple write scenarios in kv * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * updating register description for internal fw update reset wait cycle count to indicate that 5 is the minimum value allowed updating kv definition description to clarify that SHA is no longer a valid destination * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/kv_vuln_fix' with updated timestamp and hash after successful run --------- Co-authored-by: Michael Norris <[email protected]> Co-authored-by: Michael Norris <[email protected]> Co-authored-by: Michael Norris <[email protected]> Co-authored-by: Caleb <[email protected]> Co-authored-by: Caleb Whitehead <[email protected]> Co-authored-by: Kiran Upadhyayula <[email protected]>
* fix to fw upd reset signal to enable jtag during breakpoint updated predictor for fw upd reset signal * Update impactless fw update section * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/rdc_jtag_fix' with updated timestamp and hash after successful run * MICROSOFT AUTOMATED PIPELINE: Stamp 'user/dev/michnorris/rdc_jtag_fix' with updated timestamp and hash after successful run --------- Co-authored-by: Michael Norris <[email protected]> Co-authored-by: Caleb Whitehead <[email protected]>
…ist (rm kv_sha512_flow)
|
@mkurc-ant Can you take a look at the workflow issue to see if there's an easy fix? Something new going on with SCCACHE. |
calebofearth
force-pushed
the
patch_v1.0.2_lint_fixes
branch
from
0e10966 to
9e7cef1
Compare
@calebofearth I guess once this is merged you will also merge #559. Unless you want to cherry-pick the fixes from #559 here. |
@tmichalak Thanks for the suggestion. I cherry-picked those updates and reverted the SCCACHE version change. |
|
|
@nstewart-amd
@amullick007
Couldn't add you as reviewers so tagging here :)