feat: implement support for wasm data segments #23
Conversation
hir/src/module.rs
Outdated
| // If the two segments have the same size and offset, then | ||
| // if they match in all other respects, we're done. If they | ||
| // don't match, then we raise a mismatch error. | ||
| if segment.size == size { | ||
| let has_same_data = self.globals.contains_constant(&init); | ||
| if has_same_data && segment.readonly == readonly { | ||
| return Ok(()); | ||
| } | ||
| } |
There was a problem hiding this comment.
When can this situation occur? Does it correspond to two globals referring to the same data or something?
It seems strange that it would be necessary to handle this corner case here. I would expect the caller to keep track of whether a particular piece of data has already been assigned a segment, and only call declare_data_segment when a new segment is actually needed.
There was a problem hiding this comment.
I've reworked this a bit - where the code for tracking data segments lives anyway, not this specific case. Rather than having this code live as part of Module, I've moved it to a separate DataSegmentTable data structure similar to the GlobalVariableTable. I'll push those changes in the morning.
But to answer your question, we want to handle that data segments might be declared in multiple modules at once, by making sure those segments don't overlap, or if they do, that they perfectly overlap (i.e. they are duplicate declarations of the same segment). This makes it trivial to take all of the segments from multiple modules and insert them into the same segment-tracking data structure without having to account for duplicates separately.
Wasm has the concept of "data segments", essentially initializers for specific regions of memory, identified by offset from the start of linear memory, and a vector of bytes to be written starting at that offset. This is used to implement things like the `rodata` section you'd find in a typical ELF executable/library. Globals can then be exported with an address pointing into that segment, which can be useful for sharing the same data across many read-only globals, and for packing global data more efficiently. Our implementation of this is very similar, except we also allow data segments to have a size larger than the initializer (byte vector), which has the same semantics as padding the initializer with zeroes out to that size.
bitwalker
force-pushed
the
bitwalker/data-segments
branch
from
a377c23
to
4a4ce53
Compare
Wasm has the concept of "data segments", essentially initializers for specific regions of memory, identified by offset from the start of linear memory, and a vector of bytes to be written starting at that offset. This is used to implement things like the
rodatasection you'd find in a typical ELF executable/library. Globals can then be exported with an address pointing into that segment, which can be useful for sharing the same data across many read-only globals, and for packing global data more efficiently.Our implementation of this is very similar, except we also allow data segments to have a size larger than the initializer (byte vector), which has the same semantics as padding the initializer with zeroes out to that size.
--
This PR also implements formatting for segments, as well as globals, since formatting of the latter was missed in the PR which introduced them.
/cc @greenhat