Bill kindly pointed out to me that v8windbg was not handling bit_field2
correctly. The issue was that the constexpr type for ElementsKind was,
somewhat unsurprisingly, "ElementsKind", but v8windbg expected a fully-
qualified type name like "v8::internal::ElementsKind". This change
addresses the problem in two ways:
1. Update v8windbg's type resolution logic to resolve type names as if
they were used in the v8::internal namespace. This makes it more
consistent with how those type names are used in other generated
Torque code, reducing surprises and the number of times we have to
write `v8::internal::` in .tq files.
2. Add compile-time verification that any constexpr type name used as a
string in class-debug-readers-tq.cc can also resolve as a type name.
Bug: v8:9376
Change-Id: I349cd6ab586fd8345a1fa8bfc3989bb8e6376ab8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2063769
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#66633}
This reverts commit 4dc1fb4e04.
Reason for revert: the regression from the original change was likely due to unlucky factors like code alignment.
Original change's description:
> Revert "[torque] Support bitfield structs stored within Smis"
>
> This reverts commit e5e4ea962e.
>
> Reason for revert: mysterious performance regression chromium:1052756
>
> Original change's description:
> > [torque] Support bitfield structs stored within Smis
> >
> > This change moves the definition of the bits stored in DebugInfo::flags
> > to Torque, and updates the only Torque usage of that field to use more
> > natural syntax. This is intended as an example of common patterns found
> > in various other classes. Several supporting changes are required:
> >
> > 1. Add a new type representing a bitfield struct stored within a Smi. It
> > is currently called SmiTagged, but I'm open to suggestions.
> > 2. Add an enum-style output for Torque bitfield structs whose bitfields
> > occupy only one bit each.
> > 3. Add a new case to MachineOperatorReducer that makes the generated
> > code for IncBlockCounter match with what was generated before this
> > change.
> > 4. Add support for reporting these bitfields in the postmortem debugging
> > API. The format matches existing bitfields but with an offset value
> > that includes the SMI shift size.
> >
> > Bug: v8:7793
> > Change-Id: Icaecbe4a162da55d2d9a3a35a8ea85b285b2f1b7
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2028832
> > Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> > Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> > Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#66182}
>
> Bug: chromium:1052756, v8:7793
> Change-Id: I9e2897efbb6321124bf4952cf09de2f179f7310d
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2062569
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66349}
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: chromium:1052756, v8:7793
Change-Id: I6087928aa14c8551ebd294513bd8d6ffa402a0d4
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2070635
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#66465}
This reverts commit e5e4ea962e.
Reason for revert: mysterious performance regression chromium:1052756
Original change's description:
> [torque] Support bitfield structs stored within Smis
>
> This change moves the definition of the bits stored in DebugInfo::flags
> to Torque, and updates the only Torque usage of that field to use more
> natural syntax. This is intended as an example of common patterns found
> in various other classes. Several supporting changes are required:
>
> 1. Add a new type representing a bitfield struct stored within a Smi. It
> is currently called SmiTagged, but I'm open to suggestions.
> 2. Add an enum-style output for Torque bitfield structs whose bitfields
> occupy only one bit each.
> 3. Add a new case to MachineOperatorReducer that makes the generated
> code for IncBlockCounter match with what was generated before this
> change.
> 4. Add support for reporting these bitfields in the postmortem debugging
> API. The format matches existing bitfields but with an offset value
> that includes the SMI shift size.
>
> Bug: v8:7793
> Change-Id: Icaecbe4a162da55d2d9a3a35a8ea85b285b2f1b7
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2028832
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66182}
Bug: chromium:1052756, v8:7793
Change-Id: I9e2897efbb6321124bf4952cf09de2f179f7310d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2062569
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66349}
This change moves the definition of the bits stored in DebugInfo::flags
to Torque, and updates the only Torque usage of that field to use more
natural syntax. This is intended as an example of common patterns found
in various other classes. Several supporting changes are required:
1. Add a new type representing a bitfield struct stored within a Smi. It
is currently called SmiTagged, but I'm open to suggestions.
2. Add an enum-style output for Torque bitfield structs whose bitfields
occupy only one bit each.
3. Add a new case to MachineOperatorReducer that makes the generated
code for IncBlockCounter match with what was generated before this
change.
4. Add support for reporting these bitfields in the postmortem debugging
API. The format matches existing bitfields but with an offset value
that includes the SMI shift size.
Bug: v8:7793
Change-Id: Icaecbe4a162da55d2d9a3a35a8ea85b285b2f1b7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2028832
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66182}
This change adds support for the postmortem inspection library to show
the content of cached external strings if that content is available. It
also fixes a minor annoyance where strings with unavailable data would
show up as "...". Now, if fetching the very first character fails, we
omit the literal value from the output.
Bug: v8:9376
Change-Id: Id694a774c231ab3467fb59b1c149284729acfb20
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1987922
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#65961}
This change updates GetObjectProperties to list all of the bitfields
within a class field, if that class field's type is a bitfield struct.
The representation of bitfields in the GetObjectProperties response is
very similar to the representation of struct fields, but with two extra
bytes of data specifying the shift and size of the bitfield.
Bug: v8:9376
Change-Id: I40a22169f3d01652a7f2db8cface43c2a1e30cfe
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1960835
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#65610}
Part of the GetObjectProperties test case is for verifying the human-
readable brief object description string that GetObjectProperties
returns. That string might look something like this:
"xy" (0x28f038d5 <v8::internal::SeqOneByteString>)
GetObjectProperties also tries to detect known immortal objects by
recognizing their addresses, which is useful in crash dumps with limited
memory. The recognized object name, if it exists, is prepended to the
description string. In order to provide this data accurately (in builds
without pointer compression), GetObjectProperties relies on the caller
to provide the addresses of the first pages in read-only space, map
space, and old space. If the caller doesn't provide those addresses,
then GetObjectProperties does the best it can with limited information
and reports possible matches based on an object's offset within the heap
page that contains it. So the result string might look like this, if the
object happened to get allocated at a lucky offset within its page:
maybe LoadHandler3Map "xy" (0x28f038d5 <v8::internal::SeqOneByteString>)
As a result, when testing these descriptions, we should generally check
that they contain the interesting data rather than that they start with
it, because some incorrect "maybe" match with a known object might be
included at the beginning.
Bug: v8:10034
Change-Id: I0cf5afd67793a239614aba3665ef57cd2d663a47
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1950233
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#65432}
Until now, the in-object properties on JSObject have been invisible to
tools using the postmortem debugging library. With this change, those
tools will get enough information to show a flat list of property
values. This is still less powerful than the runtime printers, which can
show the corresponding key for each value, but it's a big step up from
manually inspecting memory.
This change basically requires a reimplementation of
Map::GetInObjectProperties for postmortem debugging. I'm not
enthusiastic about duplicating this logic, but it's pretty small and I
don't see any good alternatives.
As a drive-by cleanup, I moved some inline string literals into a batch
of constexpr char arrays.
Bug: v8:9376
Change-Id: Ia24c05f6e823086babaa07882d0d320ab9a225db
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1930174
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#65183}
This change defines a way that v8_debug_helper can describe object
fields which are packed structs, and uses it for the "descriptors" field
in DescriptorArray.
In more detail:
- debug-helper.h (the public interface for v8_debug_helper) adds a size
and an optional list of struct properties to ObjectProperty.
- debug-helper-internal.h mirrors those changes to the internal class
hierarchy which maintains proper unique_ptr ownership.
- In src/torque/class-debug-reader-generator.cc,
- Some existing logic is moved into smaller functions.
- New logic is added to generate the field list for structs. Example
output is included in a comment above the function
GenerateGetPropsChunkForField.
Bug: v8:9376
Change-Id: I531acac039ccb42050641448a4cbaec26186a7bc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1894362
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65079}
This change extends v8_debug_helper to export a new method that returns
a list of all known heap object types.
Why? We can substantially improve the user experience in our work-in-
progress WinDbg extension if we register handlers not only for
v8::internal::Object but for every specific HeapObject type. This has
two benefits:
- You save a click: if you're expanding a local variable of a more
specific type than Object, you can see properties immediately rather
than first needing to expand a sub-item that casts the variable to
Object.
- You retain the type hint: GetObjectProperties accepts a type hint
string, and it's super important to pass it when working in a crash
dump because the object's Map is probably inaccessible. If we have to
cast to Object first, we lose this data.
Bug: v8:9376
Change-Id: I4d635a1826574a3d08ac657e848e1fe7b83849fe
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822859
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#64331}
If we can read an object's Map pointer but not any data from the Map
itself, we may still be able to accurately describe the object's type if
the Map pointer matches one of the known Maps from the snapshot.
GetObjectProperties uses that data in one of two ways:
- If it is sure that the Map pointer matches a known Map, then it uses
the type from that Map and continues as if it read the type normally.
- If the Map pointer is at the right offset within a heap page to match
a known Map, but the caller didn't provide the addresses of the first
pages in Map space or read-only space, then the type of that Map is
just a guess and gets returned in a separate array. This gives the
caller the opportunity to present guessed types to the user, and
perhaps call again using the guessed type as the type hint.
Bug: v8:9376
Change-Id: I187f67b77e76699863a14534a9d635b79f654124
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1787986
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#63908}
v8_debug_helper attempts to flag known object pointers when it can
recognize them, even if the memory pointed to is not available in the
crash dump. In ptr-compr builds, the first pages of the map space,
read-only space, and old space are always at the same offsets within the
heap reservation region, so we can more easily detect known objects.
Bug: v8:9376
Change-Id: I04e0d2357143d753f575f556e94f8fd42ce9d811
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1783729
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#63624}
This change provides a quick way to see string contents in postmortem
debugging sessions, without digging through a (possibly very large, in
the case of ConsString) tree of properties. As well as being convenient
for inspecting String objects, this functionality will also be necessary
for displaying property names on JSReceiver objects. In order to support
custom behaviors for specific classes, this change extends the existing
generated debug reader classes with a visitor pattern.
Bug: v8:9376
Change-Id: I70eab9ea4e74ca0fab39bf5998d6a602716a4202
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1771939
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#63485}
This change adds the indexed field for the characters in the definition
of sequential string types, and introduces support for recognizing the
various specific string types in v8_debug_helper. In an attempt to
avoid duplicating info about string instance types, it also refactors
String::Get so that StringShape (a simple class usable by postmortem
tools) can dispatch using a class that defines behaviors for each
concrete type.
Bug: v8:9376
Change-Id: Id0653040f6decddc004c73f8fe93d2187828c2c6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1735795
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Cr-Commit-Position: refs/heads/master@{#63352}
This is a reland of 517ab73fd7
Updates since original: now compressed pointers passed to the function
GetObjectProperties are required to be sign-extended. Previously, the
function allowed zero-extended values, but that led to ambiguity on
pointers like 0x88044919: is it compressed or is the heap range actually
centered on 0x100000000?
Original change's description:
> Add postmortem debugging helper library
>
> This change begins to implement the functionality described in
> https://docs.google.com/document/d/1evHnb1uLlSbvHAAsmOXyc25x3uh1DjgNa8u1RHvwVhk/edit#
> for investigating V8 state in crash dumps.
>
> This change adds a new library, v8_debug_helper, for providing platform-
> agnostic assistance with postmortem debugging. This library can be used
> by extensions built for debuggers such as WinDbg or lldb. Its public API
> is described by debug-helper.h; currently the only method it exposes is
> GetObjectProperties, but we'd like to add more functionality over time.
> The API surface is restricted to plain C-style structs and pointers, so
> that it's easy to link from a debugger extension built with a different
> toolchain.
>
> This change also adds a new cctest file to exercise some basic
> interaction with the new library.
>
> The API function GetObjectProperties takes an object pointer (which
> could be compressed, or weak, or a SMI), and returns a string
> description of the object and a list of properties the object contains.
> For now, the list of properties is entirely based on Torque object
> definitions, but we expect to add custom properties in future updates so
> that it can be easier to make sense of complex data structures such as
> dictionaries.
>
> GetObjectProperties does several things that are intended to generate
> somewhat useful results even in cases where memory may be corrupt or
> unavailable:
> - The caller may optionally provide a type string which will be used if
> the memory for the object's Map is inaccessible.
> - All object pointers are compared against the list of known objects
> generated by mkgrokdump. The caller may optionally provide the
> pointers for the first pages of various heap spaces, to avoid spurious
> matches. If those pointers are not provided, then any matches are
> prefixed with "maybe" in the resulting description string, such as
> "maybe UndefinedValue (0x4288000341 <Oddball>)".
>
> Bug: v8:9376
>
> Change-Id: Iebf3cc2dea3133c7811bcefcdf38d9458b02fded
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1628012
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Reviewed-by: Michael Stanton <mvstanton@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#62882}
Bug: v8:9376
Change-Id: I866a1cc9d4c34bfe10c7b98462451fe69763cf3f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1717090
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#63008}
This reverts commit 517ab73fd7.
Reason for revert: Test failures https://bugs.chromium.org/p/v8/issues/detail?id=9538
Original change's description:
> Add postmortem debugging helper library
>
> This change begins to implement the functionality described in
> https://docs.google.com/document/d/1evHnb1uLlSbvHAAsmOXyc25x3uh1DjgNa8u1RHvwVhk/edit#
> for investigating V8 state in crash dumps.
>
> This change adds a new library, v8_debug_helper, for providing platform-
> agnostic assistance with postmortem debugging. This library can be used
> by extensions built for debuggers such as WinDbg or lldb. Its public API
> is described by debug-helper.h; currently the only method it exposes is
> GetObjectProperties, but we'd like to add more functionality over time.
> The API surface is restricted to plain C-style structs and pointers, so
> that it's easy to link from a debugger extension built with a different
> toolchain.
>
> This change also adds a new cctest file to exercise some basic
> interaction with the new library.
>
> The API function GetObjectProperties takes an object pointer (which
> could be compressed, or weak, or a SMI), and returns a string
> description of the object and a list of properties the object contains.
> For now, the list of properties is entirely based on Torque object
> definitions, but we expect to add custom properties in future updates so
> that it can be easier to make sense of complex data structures such as
> dictionaries.
>
> GetObjectProperties does several things that are intended to generate
> somewhat useful results even in cases where memory may be corrupt or
> unavailable:
> - The caller may optionally provide a type string which will be used if
> the memory for the object's Map is inaccessible.
> - All object pointers are compared against the list of known objects
> generated by mkgrokdump. The caller may optionally provide the
> pointers for the first pages of various heap spaces, to avoid spurious
> matches. If those pointers are not provided, then any matches are
> prefixed with "maybe" in the resulting description string, such as
> "maybe UndefinedValue (0x4288000341 <Oddball>)".
>
> Bug: v8:9376
>
> Change-Id: Iebf3cc2dea3133c7811bcefcdf38d9458b02fded
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1628012
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Reviewed-by: Michael Stanton <mvstanton@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#62882}
TBR=yangguo@chromium.org,mvstanton@chromium.org,jgruber@chromium.org,tebbi@chromium.org,seth.brenith@microsoft.com
Change-Id: Ia078f2e8d101d2375b5db88021b2d65d28f1b075
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:9376
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1716033
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62899}
This change begins to implement the functionality described in
https://docs.google.com/document/d/1evHnb1uLlSbvHAAsmOXyc25x3uh1DjgNa8u1RHvwVhk/edit#
for investigating V8 state in crash dumps.
This change adds a new library, v8_debug_helper, for providing platform-
agnostic assistance with postmortem debugging. This library can be used
by extensions built for debuggers such as WinDbg or lldb. Its public API
is described by debug-helper.h; currently the only method it exposes is
GetObjectProperties, but we'd like to add more functionality over time.
The API surface is restricted to plain C-style structs and pointers, so
that it's easy to link from a debugger extension built with a different
toolchain.
This change also adds a new cctest file to exercise some basic
interaction with the new library.
The API function GetObjectProperties takes an object pointer (which
could be compressed, or weak, or a SMI), and returns a string
description of the object and a list of properties the object contains.
For now, the list of properties is entirely based on Torque object
definitions, but we expect to add custom properties in future updates so
that it can be easier to make sense of complex data structures such as
dictionaries.
GetObjectProperties does several things that are intended to generate
somewhat useful results even in cases where memory may be corrupt or
unavailable:
- The caller may optionally provide a type string which will be used if
the memory for the object's Map is inaccessible.
- All object pointers are compared against the list of known objects
generated by mkgrokdump. The caller may optionally provide the
pointers for the first pages of various heap spaces, to avoid spurious
matches. If those pointers are not provided, then any matches are
prefixed with "maybe" in the resulting description string, such as
"maybe UndefinedValue (0x4288000341 <Oddball>)".
Bug: v8:9376
Change-Id: Iebf3cc2dea3133c7811bcefcdf38d9458b02fded
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1628012
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62882}
