This fix was created after profiling the target creation of a large C/C++/ObjC application that contained almost 4,000,000 redacted symbol names. The symbol table parsing code was creating names for each of these synthetic symbols and adding them to the name indexes. The code was also adding the object file basename to the end of the symbol name which doesn't allow symbols from different shared libraries to share the names in the constant string pool.
Prior to this fix this was creating 180MB of "___lldb_unnamed_symbol" symbol names and was taking a long time to generate each name, add them to the string pool and then add each of these names to the name index.
This patch fixes the issue by:
- not adding a name to synthetic symbols at creation time, and allows name to be dynamically generated when accessed
- doesn't add synthetic symbol names to the name indexes, but catches this special case as name lookup time. Users won't typically set breakpoints or lookup these synthetic names, but support was added to do the lookup in case it does happen
- removes the object file baseanme from the generated names to allow the names to be shared in the constant string pool
Prior to this fix the startup times for a large application was:
12.5 seconds (cold file caches)
8.5 seconds (warm file caches)
After this fix:
9.7 seconds (cold file caches)
5.7 seconds (warm file caches)
The names of the symbols are auto generated by appending the symbol's UserID to the end of the "___lldb_unnamed_symbol" string and is only done when the name is requested from a synthetic symbol if it has no name.
Differential Revision: https://reviews.llvm.org/D105160
Reverts commits:
"Fix failing tests after https://reviews.llvm.org/D104488."
"Fix buildbot failure after https://reviews.llvm.org/D104488."
"Create synthetic symbol names on demand to improve memory consumption and startup times."
This series of commits broke the windows lldb bot and then failed to fix all of the failing tests.
In D98289#inline-939112 @dblaikie said:
Perhaps this could be more informative about what makes the range list
index of 0 invalid? "index 0 out of range of range list table (with
range list base 0xXXX) with offset entry count of XX (valid indexes
0-(XX-1))" Maybe that's too verbose/not worth worrying about since
this'll only be relevant to DWARF producers trying to debug their
DWARFv5, maybe no one will ever see this message in practice. Just
a thought.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D102851
DW_AT_ranges can use DW_FORM_sec_offset (instead of DW_FORM_rnglistx).
In such case DW_AT_rnglists_base does not need to be present.
DWARF-5 spec:
"If the offset_entry_count is zero, then DW_FORM_rnglistx cannot
be used to access a range list; DW_FORM_sec_offset must be used
instead. If the offset_entry_count is non-zero, then
DW_FORM_rnglistx may be used to access a range list;"
This fix is for TestTypeCompletion.py category `dwarf` using GCC with DWARF-5.
The fix just provides GetRnglist() lazy getter for `m_rnglist_table`.
The testcase is easier to review by:
diff -u lldb/test/Shell/SymbolFile/DWARF/DW_AT_low_pc-addrx.s \
lldb/test/Shell/SymbolFile/DWARF/DW_AT_range-DW_FORM_sec_offset.s
Differential Revision: https://reviews.llvm.org/D98289
DWARF allows .dwo file paths to be relative rather than absolute. When
they are relative, DWARF uses DW_AT_comp_dir to find the .dwo
file. DW_AT_comp_dir can also be relative, making the entire search
patch for the .dwo file relative. In this case, LLDB currently
searches relative to its current working directory, i.e. the directory
from which the debugger was launched. This is not right, as the
compiler, which generated the relative paths, can have no idea where
the debugger will be launched. The correct thing is to search relative
to the location of the executable binary. That is what this patch
does.
Differential Revision: https://reviews.llvm.org/D97786
DWARF allows .dwo file paths to be relative rather than absolute. When
they are relative, DWARF uses DW_AT_comp_dir to find the .dwo
file. DW_AT_comp_dir can also be relative, making the entire search
patch for the .dwo file relative. In this case, LLDB currently
searches relative to its current working directory, i.e. the directory
from which the debugger was launched. This is not right, as the
compiler, which generated the relative paths, can have no idea where
the debugger will be launched. The correct thing is to search relative
to the location of the executable binary. That is what this patch
does.
Differential Revision: https://reviews.llvm.org/D97786
These tests fail if you build without the x86 llvm backend.
Either because they use an x86 triple or try to backtrace which
requires some x86 knowledge to see all frames.
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D100194
By moving them into a folder with a local lit config
requiring x86. All these tests use x86 target triples.
There are two tests that require target-x86_64 because
they run program files (instead of just needing the backend).
Those are moved to the x86 folder also but their REQUIRES are
unchanged.
Reviewed By: JDevlieghere
Differential Revision: https://reviews.llvm.org/D100193
The file contained bogus input - the DIE list was not properly
terminated. This should not cause a crash, but it seems it was crashing
at least on linux arm and x86 windows.
SymbolFileDWARF::ResolveSymbolContext is currently unaware that in DWARF5 the primary file is specified at file index 0. As a result it misses to correctly resolve the symbol context for the primary file when DWARF5 debug data is used and the primary file is only specified at index 0.
This change makes use of CompileUnit::ResolveSymbolContext to resolve the symbol context. The ResolveSymbolContext in CompileUnit has been previously already updated to reflect changes in DWARF5
and contains a more readable version. It can resolve more, but will also do a bit more work than
SymbolFileDWARF::ResolveSymbolContext (getting the Module, and going through SymbolFileDWARF::ResolveSymbolContextForAddress), however, it's mostly directed by $resolve_scope
what will be resolved, and ensures that code is easier to maintain if there's only one path.
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D98619
Apply changes from https://reviews.llvm.org/D91014 to other places where DWARF entries are being processed.
Test case is provided by @jankratochvil.
The test is marked to run only on x64 and exclude Windows and Darwin, because the assembly is not OS-independent.
(First attempt https://reviews.llvm.org/D96778 broke the build bots)
Reviewed By: jankratochvil
Differential Revision: https://reviews.llvm.org/D97765
In DWARF v4 compile units go in .debug_info and type units go in
.debug_types. However, in v5 both kinds of units are in .debug_info.
Therefore we can't decide whether to use the CU or TU index just by
looking at which section we're reading from. We have to wait until we
have read the unit type from the header.
Differential Revision: https://reviews.llvm.org/D96194
Currently TypePrinter lumps anonymous classes and unnamed classes in one group "anonymous" this is not correct and can be confusing in some contexts.
Differential Revision: https://reviews.llvm.org/D96807
Currently TypePrinter lumps anonymous classes and unnamed classes in one group "anonymous" this is not correct and can be confusing in some contexts.
Differential Revision: https://reviews.llvm.org/D96807
Finishing out the support (to the best of my knowledge/based on current
testing running the whole check-lldb with a clang forcibly using
DW_AT_ranges on all DW_TAG_subprograms) for this feature.
Differential Revision: https://reviews.llvm.org/D94064
Local values are constants or addresses that can't be folded into
the instruction that uses them. FastISel materializes these in a
"local value" area that always dominates the current insertion
point, to try to avoid materializing these values more than once
(per block).
https://reviews.llvm.org/D43093 added code to sink these local
value instructions to their first use, which has two beneficial
effects. One, it is likely to avoid some unnecessary spills and
reloads; two, it allows us to attach the debug location of the
user to the local value instruction. The latter effect can
improve the debugging experience for debuggers with a "set next
statement" feature, such as the Visual Studio debugger and PS4
debugger, because instructions to set up constants for a given
statement will be associated with the appropriate source line.
There are also some constants (primarily addresses) that could be
produced by no-op casts or GEP instructions; the main difference
from "local value" instructions is that these are values from
separate IR instructions, and therefore could have multiple users
across multiple basic blocks. D43093 avoided sinking these, even
though they were emitted to the same "local value" area as the
other instructions. The patch comment for D43093 states:
Local values may also be used by no-op casts, which adds the
register to the RegFixups table. Without reversing the RegFixups
map direction, we don't have enough information to sink these
instructions.
This patch undoes most of D43093, and instead flushes the local
value map after(*) every IR instruction, using that instruction's
debug location. This avoids sometimes incorrect locations used
previously, and emits instructions in a more natural order.
In addition, constants materialized due to PHI instructions are
not assigned a debug location immediately; instead, when the
local value map is flushed, if the first local value instruction
has no debug location, it is given the same location as the
first non-local-value-map instruction. This prevents PHIs
from introducing unattributed instructions, which would either
be implicitly attributed to the location for the preceding IR
instruction, or given line 0 if they are at the beginning of
a machine basic block. Neither of those consequences is good
for debugging.
This does mean materialized values are not re-used across IR
instruction boundaries; however, only about 5% of those values
were reused in an experimental self-build of clang.
(*) Actually, just prior to the next instruction. It seems like
it would be cleaner the other way, but I was having trouble
getting that to work.
This reapplies commits cf1c774d and dc35368c, and adds the
modification to PHI handling, which should avoid problems
with debugging under gdb.
Differential Revision: https://reviews.llvm.org/D91734
gcc already produces debug info with this form
-freorder-block-and-partition
clang produces this sort of thing with -fbasic-block-sections and with a
coming-soon tweak to use ranges in DWARFv5 where they can allow greater
reuse of debug_addr than the low/high_pc forms.
This fixes the case of breaking on a function name, but leaves broken
printing a variable - a follow-up commit will add that and improve the
test case to match.
Differential Revision: https://reviews.llvm.org/D94063
In split DWARF v5 files, the DWO id is no longer in the DW_AT_GNU_dwo_id
attribute. It's in the CU header instead. This change makes lldb look in
both places.
Differential Revision: https://reviews.llvm.org/D93444
This reverts commit cf1c774d6a.
This change caused several regressions in the gdb test suite - at least
a sample of which was due to line zero instructions making breakpoints
un-lined. I think they're worth investigating/understanding more (&
possibly addressing) before moving forward with this change.
Revert "[FastISel] NFC: Clean up unnecessary bookkeeping"
This reverts commit 3fd39d3694.
Revert "[FastISel] NFC: Remove obsolete -fast-isel-sink-local-values option"
This reverts commit a474657e30.
Revert "Remove static function unused after cf1c774."
This reverts commit dc35368ccf.
Revert "[lldb] Fix TestThreadStepOut.py after "Flush local value map on every instruction""
This reverts commit 53a14a47ee.
Local values are constants or addresses that can't be folded into
the instruction that uses them. FastISel materializes these in a
"local value" area that always dominates the current insertion
point, to try to avoid materializing these values more than once
(per block).
https://reviews.llvm.org/D43093 added code to sink these local
value instructions to their first use, which has two beneficial
effects. One, it is likely to avoid some unnecessary spills and
reloads; two, it allows us to attach the debug location of the
user to the local value instruction. The latter effect can
improve the debugging experience for debuggers with a "set next
statement" feature, such as the Visual Studio debugger and PS4
debugger, because instructions to set up constants for a given
statement will be associated with the appropriate source line.
There are also some constants (primarily addresses) that could be
produced by no-op casts or GEP instructions; the main difference
from "local value" instructions is that these are values from
separate IR instructions, and therefore could have multiple users
across multiple basic blocks. D43093 avoided sinking these, even
though they were emitted to the same "local value" area as the
other instructions. The patch comment for D43093 states:
Local values may also be used by no-op casts, which adds the
register to the RegFixups table. Without reversing the RegFixups
map direction, we don't have enough information to sink these
instructions.
This patch undoes most of D43093, and instead flushes the local
value map after(*) every IR instruction, using that instruction's
debug location. This avoids sometimes incorrect locations used
previously, and emits instructions in a more natural order.
This does mean materialized values are not re-used across IR
instruction boundaries; however, only about 5% of those values
were reused in an experimental self-build of clang.
(*) Actually, just prior to the next instruction. It seems like
it would be cleaner the other way, but I was having trouble
getting that to work.
Differential Revision: https://reviews.llvm.org/D91734
I found a few cases where entries in the debug_line for a specific line of code have invalid entries (the address is outside of a code section or no section at all) and also valid entries. When this happens lldb might not set the breakpoint because the first line entry it will find in the line table might be the invalid one and since it's range is "invalid" no location is resolved. To get around this I changed the way we parse the line sequences to ignore those starting at an address under the first code segment.
Greg suggested to implement it this way so we don't need to check all sections for every line sequence.
Reviewed By: clayborg
Differential Revision: https://reviews.llvm.org/D87172
This would be reproducible in future DWZ category of the testsuite as:
Failed Tests (1):
lldb-api :: python_api/symbol-context/two-files/TestSymbolContextTwoFiles.py
Differential Revision: https://reviews.llvm.org/D91014
