AArch64 codec is not scalable #5326
Assignees
Comments
|
Would this be related to #4393? |
Yes. Partitioning by version will cleanly separate the current method of definition and generation from a future method. |
AssadHashmi
added a commit
that referenced
this issue
Feb 8, 2022
The current methods of defining and generating encode and decode functionality are not scalable for implementing support of ISA versions after v8.0 This patch splits up the codec.txt file into codec_<version>.txt files and opnd_defs.txt. Generated code is now split out into ISA version specific files. Tests and related files have not been split up yet. That will happen as part of new ISA version implementations. Issue: #5326, #4393
AssadHashmi
added a commit
that referenced
this issue
Feb 8, 2022
The current methods of defining and generating encode and decode functionality are not scalable for implementing support of ISA versions after v8.0 This patch splits up the codec.txt file into codec_<version>.txt files and opnd_defs.txt. Generated code is now split out into ISA version specific files. Tests and related files have not been split up yet. That will happen as part of new ISA version implementations. Issue: #5326, #4393
|
Is there a plan to expose the ISA version in the IR interface? Would it be in the instruction structure, or all ISA version changes are assumed to be completely isolated by opcode (i.e., a version never adds a variant of an existing opcode) and it can be looked up purely from the opcode? |
joshua-warburton
added a commit
that referenced
this issue
Jul 11, 2022
Previously codecsort.py was only applied to a single codec at a time and thus couldn't allocate enums uniquely as it didn't know about the enum values in other files. This patch simplifies the process of calling codecsort.py so that it can consider all the codecs at once, and thus rewrite them as needed. It also changes the default mode so that it returns an exit code of 1 if it detects a rewrite is needed making checking th order easier. Issue: #5326 Change-Id: Id433cb9123e7cf901ae0e6bf14e84590cf39eadb
joshua-warburton
added a commit
that referenced
this issue
Jul 12, 2022
…5558) Previously codecsort.py was only applied to a single codec at a time and thus couldn't allocate enums uniquely as it didn't know about the enum values in other files. This patch simplifies the process of calling codecsort.py so that it can consider all the codecs at once, and thus rewrite them as needed. It also changes the default mode so that it returns an exit code of 1 if it detects a rewrite is needed making checking the order easier. Issue: #5326
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The current methods of defining and generating encode and decode functionality are not scalable for implementing support of ISA versions after v8.0
At present AArch64 has one
codec.txtand onecodec.cfile which defines all instructions currently supported as well as all the bitmask and operand definitions. Encode and decode data and code generated fromcodec.txtare also all in one set of files regardless of ISA version.Test files like
dis-a64.txtandir_aarch64.ccontain all encode/decode instructions tests.This issue will track separation of the codec by ISA version for the purposes of reduced maintenance overhead and more reliable functionality.
The text was updated successfully, but these errors were encountered: