Equate fieldset names #33
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Florob
reviewed
Jun 4, 2023
Omitting this check allowed the combine pass to eagerly merge fieldsets it believed were equivalent. Consider the LPI2C.SIER register. On the 1000 series, the "address match 1 interrupt enable" field is called "AM1IE." And on the 1100 series, the same field is called "AM1F." Same meaning, but a different name. By eagerly combining, we have no need to generate another lpi2c.rs files, and we have 1675 fewer lines of code. The assumption fails when that name difference is meaningful, like when NXP changes the CCM.CS1CDR field at offset 25 from FLEXIO1_CLK_PODF to FLEXIO2_CLK_PODF. Those aren't equivalent fields, but we don't know that because we're not checking the field name. It's always correct to not combine. And turns out it's not adding too much code; we can patch the small name corrections like the LPI2C example. Those patch commits are children of this commit.
See the parent commit for more context. With this patch, we can still combine the LPI2C peripherals across all peripherals. Since field combinations wasn't considering names, the combiner was free to select any version of the field as the base. It happened to pick the "less good" field name for all peripherals. This commit ends up changing the source as a consequence of the rename.
Selecting the longer name for clarity, and because the longer name is used in more MCU SVDs.
This fixes the CCM.CS1CDR FLEXIO1 vs FLEXIO2 field name descrepancy. It also seems to correct a field name in the an 1176 SNVS register. The newly-added TRNG is because there's a field name that correctly changes as a result of the combiner policy change.
mciantyre
force-pushed
the
equate-fieldset-names
branch
from
2276d76 to
2a43e2d
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Omitting checks for fieldset names allowed the combine pass to eagerly merge fieldsets it believed were equivalent. Consider the LPI2C.SIER register. On the 1000 series, the "address match 1 interrupt enable" field is called "AM1IE." And on the 1100 series, the same field is called "AM1F." Same meaning, but a different name. By eagerly combining, we have no need to generate another
lpi2c.rsfile, and we have 1675 fewer lines of code.The assumption fails when that name difference is meaningful, like when NXP changes the CCM.CS1CDR field at offset 25 from FLEXIO1_CLK_PODF to FLEXIO2_CLK_PODF. Those aren't equivalent fields, but we don't know that because we're not checking the field name.
It's always correct to not combine. And turns out it's not adding too much code; we can patch the small name corrections like the LPI2C example.
This PR ensures the combiner always equates fieldset names. It also adds raltool patches to rename specific fieldsets that truly should have the same name. The PR corrects the concrete case identified in #32, but it does not fully address the greediness problem. See individual commits for more information and to simplify review.