Fix stack overflow when generating debuginfo for 'recursive' type #59446
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
|
r? @cramertj (rust_highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
|
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
Aaron1011
force-pushed
the
fix/debuginfo-overflow
branch
from
f54b7eb
to
c1b9632
Compare
oli-obk
reviewed
Mar 28, 2019
| // This kind of type cannot be properly represented | ||
| // via LLVM debuginfo. As a workaround, | ||
| // we register a temporary Ty to metadata mapping | ||
| // for the function before we compute its actual metadat.a |
There was a problem hiding this comment.
Suggested change
| // for the function before we compute its actual metadat.a | |
| // for the function before we compute its actual metadata |
| size.bits(), | ||
| align.bits() as u32, | ||
| DW_ATE_unsigned) | ||
|
|
There was a problem hiding this comment.
there are a bunch of empty lines here
| // recursing forever. | ||
| // | ||
| // This function is used to remove the temporary metadata | ||
| // mapping after we've computed the actual metadat |
There was a problem hiding this comment.
Suggested change
| // mapping after we've computed the actual metadat | |
| // mapping after we've computed the actual metadata |
| t.fn_sig(cx.tcx), | ||
| usage_site_span).metadata; | ||
|
|
||
| type_map.borrow_mut().remove_type(t); |
There was a problem hiding this comment.
why are we registering this only temporarily and don't just leave it in?
There was a problem hiding this comment.
The temporary metadata is only used in the case where we recurse back to processing the same type.
The actual type will be inserted into the map at the bottom of the function.
|
@oli-obk I've made the changes you requested |
|
@bors r+ |
|
📌 Commit 50ab958 has been approved by |
bors
added
S-waiting-on-bors
Centril
added a commit
to Centril/rust
that referenced
this pull request
Mar 29, 2019
…=oli-obk
Fix stack overflow when generating debuginfo for 'recursive' type
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes rust-lang#58463
|
Speculatively assigning blame of failure in #59538 (comment), @bors r- |
bors
added
S-waiting-on-author
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes rust-lang#58463
Aaron1011
force-pushed
the
fix/debuginfo-overflow
branch
from
50ab958
to
d2584e3
Compare
|
@Centril: Fixed |
Centril
added
S-waiting-on-review
pnkfelix
reviewed
Apr 1, 2019
| // (e.g. MyType<fn() -> u8, fn() -> u8> | ||
| // | ||
| // We only care about avoiding recursing | ||
| // directly back to the type we're currentlu |
There was a problem hiding this comment.
typo: "currentlu" should be "currently"
|
@pnkfelix Fixed |
|
r? @oli-obk |
|
📌 Commit c13daeb has been approved by |
bors
added
S-waiting-on-bors
Centril
added a commit
to Centril/rust
that referenced
this pull request
Apr 2, 2019
…=oli-obk
Fix stack overflow when generating debuginfo for 'recursive' type
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes rust-lang#58463
Centril
added a commit
to Centril/rust
that referenced
this pull request
Apr 2, 2019
…=oli-obk
Fix stack overflow when generating debuginfo for 'recursive' type
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes rust-lang#58463
bors
added a commit
that referenced
this pull request
Apr 2, 2019
Rollup of 4 pull requests Successful merges: - #59166 (resolve: collect trait aliases along with traits) - #59341 (Fix custom relative libdir) - #59446 (Fix stack overflow when generating debuginfo for 'recursive' type) - #59529 (Added documentation on the remainder (Rem) operator for floating points.) Failed merges: r? @ghost
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.
By using 'impl trait', it's possible to create a self-referential
type as follows:
fn foo() -> impl Copy { foo }
This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...
Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run
However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.
To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.
To ensure that we emit something, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful
Fixes #58463