Stream: git-wasmtime

Topic: wasmtime / PR #2276 [debug] Normalise value prior to righ...


view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:31):

ggreif opened PR #2276 from main to main:

by first left-aligning (shift left by 32 bits)
then shifting back (respecting signedness)

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:31):

ggreif edited PR #2276 from main to main:

by first left-aligning (shift left by 32 bits)
then shifting back (respecting signedness)

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:34):

bjorn3 created PR Review Comment:

                   // Extend with the (arithmetic) shift.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:34):

bjorn3 submitted PR Review.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:34):

ggreif updated PR #2276 from main to main:

by first left-aligning (shift left by 32 bits)
then shifting back (respecting signedness)

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:39):

ggreif submitted PR Review.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:39):

ggreif created PR Review Comment:

                    // Extend with the (arithmetic) shift.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:39):

ggreif updated PR #2276 from main to main:

by first left-aligning (shift left by 32 bits)
then shifting back (respecting signedness)

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:41):

ggreif submitted PR Review.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:41):

ggreif created PR Review Comment:

                Operation::Shr { .. } | Operation::Shra { .. } => {

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:42):

ggreif updated PR #2276 from main to main:

by first left-aligning (shift left by 32 bits)
then shifting back (respecting signedness)

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 12:55):

ggreif edited PR #2276 from main to main:

Due to the 32-bit WASM/64-bit wasmtime VM schism the DW_OP_shr* operations end up with random bits frobbed from the upper portion of the 64-bit operand word. I plan to submit a fix to the location translator that inserts (something akin to) DW_OP_lit32; DW_OP_shl; DW_OP_lit32; DW_OP_shrX in front of DW_OP_shrX. The rationale is that this will clean out the upper portion, respecting 32-bit signedness.

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 13:07):

ggreif updated PR #2276 from main to main:

Due to the 32-bit WASM/64-bit wasmtime VM schism the DW_OP_shr* operations end up with random bits frobbed from the upper portion of the 64-bit operand word. I plan to submit a fix to the location translator that inserts (something akin to) DW_OP_lit32; DW_OP_shl; DW_OP_lit32; DW_OP_shrX in front of DW_OP_shrX. The rationale is that this will clean out the upper portion, respecting 32-bit signedness.

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 13:23):

ggreif updated PR #2276 from main to main:

Due to the 32-bit WASM/64-bit wasmtime VM schism the DW_OP_shr* operations end up with random bits frobbed from the upper portion of the 64-bit operand word. I plan to submit a fix to the location translator that inserts (something akin to) DW_OP_lit32; DW_OP_shl; DW_OP_lit32; DW_OP_shrX in front of DW_OP_shrX. The rationale is that this will clean out the upper portion, respecting 32-bit signedness.

TODO:

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 13:26):

ggreif edited PR #2276 from main to main:

Due to the 32-bit WASM/64-bit wasmtime VM schism the DW_OP_shr* operations end up with random bits frobbed from the upper portion of the 64-bit operand word. I plan to submit a fix to the location translator that inserts (something akin to) DW_OP_lit32; DW_OP_shl; DW_OP_lit32; DW_OP_shrX in front of DW_OP_shrX. The rationale is that this will clean out the upper portion, respecting 32-bit signedness.

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 13:32):

ggreif has marked PR #2276 as ready for review.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 07 2020 at 13:33):

ggreif edited PR #2276 from main to main:

Due to the 32-bit WASM/64-bit wasmtime VM schism the DW_OP_shr* operations end up with random bits frobbed from the upper portion of the 64-bit operand word. This fixes up to the location translator such that it inserts (something akin to) DW_OP_lit32; DW_OP_shl; DW_OP_lit32; DW_OP_shrX in front of DW_OP_shrX. The rationale is that this will clean out the upper portion, respecting 32-bit signedness.

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

Please ensure all communication adheres to the code of conduct.
-->


Last updated: Dec 23 2024 at 12:05 UTC