On a big executable, the trunk needs:
dinfo: 134873088/71438336 max/curr mmap'd, 134607808/66717872 max/curr
With the patch, we have:
dinfo: 99065856/56836096 max/curr mmap'd, 97883776/51663656 max/curr
So, peak dinfo memory decreases by about 36Mb, and final by 15Mb.
(for info, valgrind 3.9.0 uses
dinfo: 158941184/109666304 max/curr mmap'd, 156775944/107590656 max/curr
So, compared to 3.9.0, dinfo peak decreases by about 40%, and the final
memory is divided by more than 2).
The memory decrease is obtained by:
* using a dedup pool to store filename/dirname pair for the loctab source/line
information.
As typically, there is not a lot of such pairs, typically a UShort is
good enough to identify a fn/dn pair in a dedup pool.
To avoid losing memory due to alignment, the fndn indexes are stored
in a "parallel" array to the DiLoc loctab array, with entries having
1, or 2 or 4 bytes according to the nr of fn/dn pairs in the dedup pool.
See priv_storage.h comments for details.
(there was a extensible WordArray local implementation in readdwarf.c.
As with this change, we use an xarray, the local implementation was
removed).
* the memory needed for --read-inline-info is slightly decreased (-2Mb)
by removing the (unused) dirname from the DiInlLoc struct.
Handling dirname for inlined function caller implies to rework
the dwarf3 parser read_filename_table common to the var and inlinfo parser.
Waiting for this to be done, the dirname component is removed from DiInlLoc.
* the stabs reader (readstabs.c) is broken since 3.9.0.
For this change, the code has been updated to make it compile with the new
DiLoc/FnDn dedup pool. As the code is completely broken, a vg_assert(0)
has been put at the begin of the stabs reader.
* the pdb reader (readpdb.c) has been trivially updated and should still work.
It has not been tested (how do we test this ?).
A follow-up patch will be done to avoid doing too many calls to
ML_(addFnDn) : instead of having one call per ML_(addLineInfo), one
should have a single call done when reading the filename table.
This has also be tested in an outer/inner setup, to verify no
memory leak/bugs.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@14158
On a big executable, the trunk needs:
dinfo: 155844608/106737664 max/curr mmap'd 155572624/102276760 max/curr
With the patch, we have:
dinfo: 134873088/70389760 max/curr mmap'd 134607808/66717512 max/curr
So, peak dinfo memory decreases by 21Mb, and final by 36Mb.
The memory decrease is obtained by:
* using a dedup pool to store the machine dependent part (cfsi_m)
of the cfsi information as this information is highly duplicated.
For x86 and arm64, the duplication factor of cfsi machine dependent
part is very high (up to a factor 60).
For arm64, it is more like a factor 3.
A 'variable size' (1, 2 or 4 bytes) is automatically used to identify
the cfsi_m, if there is less than or more than 255/64K different cfsi_m.
* not storing explicitely the length of a range for which a cfsi_m
is to be used: in a large majority of the cases, ranges are
consecutive, and so the end of a range is just one byte before
the start of the next range.
So, we do not store the length of the ranges.
If there is a hole between 2 ranges, the hole is stored explicitely
as a range in which we have no cfsi_m information.
On x86 and amd64, we have quite some holes (something like one hole
every 7 cfsi). On arm64, we have very few holes (less than one hole
every 50 cfsi).
Even with the nr of holes on x86/amd64, it is more memory efficient
to store the holes rather than to store the length of each cfsi.
* Merging consecutive ranges that have the same cfsi_m info:
Many cfsi are "mergeable": there is no hole between 2 cfsi, and their
machine dependent part is identical
(I guess the unwind info needed by valgrind is subset of the full
unwind info, and so, the cfsi entries are not merged by the compiler,
but can be merged for simple unwind). Depending on the platform
(x86, amd64, arm64) and of the library/object file, we can have a
significant nr of mergeable entries.
The patch is not very small, but a lot is mechanical changes.
The patch has been compiled and tested on x86/amd64/ppc32/ppc64
(but ppc does not use cfsi so that just verifies it compiles).
It has been compiled on arm64, and "tested" by launching valgrind on
one executable.
It has not been compiled on s390 and mips.
With some luck, maybe it will compile on these platforms.
And if that uses the whole provision of luck for 2014, it might even work
on these platforms :).
If it does not compile, the fix should be straightforward.
Runtime problems might be more tricky (but arm64 "worked out of the box"
once x86/amd64 were ok).
This has also be tested in an outer/inner setup, to verify no memory leak/bugs.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@14129
Reading header length and values in external line info was incorrect at
some places as it used offsets based on dw64 that came from .debug_info.
Instead, offsets should be calculated based on is64 from .debug_line.
This issue surfaced in MIPS64 port, and it was discussed at:
https://bugs.kde.org/show_bug.cgi?id=313267#c20
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@13373
Bug #305513. We should only read the first DIE of a compilation unit.
Each compilation unit header is followed by a single DW_TAG_compile_unit
(or DW_TAG_partial_unit, but those aren't important here) and its children.
There is no reason to read any of the children at this point. If the first
DIE isn't a DW_TAG_compile_unit we are done, none of the child DIEs will
provide any useful information.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@13369
Bug #305513 contained a patch for some extra robustness checks. But
the real cause of crashing in the read_unitinfo_dwarf2 DWARF reader
seemed to have been this issue where DWARF version 2 DWZ partial_units
were read and DW_FORM_ref_addr had an unexpected size. This combination
is rare. DWARF version 4 is the current default version of GCC.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@13367
Necessary changes to Valgrind to support MIPS64LE on Linux.
Minor cleanup/style changes embedded in the patch as well.
The change corresponds to r2687 in VEX.
Patch written by Dejan Jevtic and Petar Jovanovic.
More information about this issue:
https://bugs.kde.org/show_bug.cgi?id=313267
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@13292
functions that do FP arithmetic. This is due to the Dwarf3 CFI
mentioning Dwarf registers above N_CFI_REGS, in particular FP
registers, which have values of about 80. This fixes the problem by
increasing N_CFI_REGS to a level that covers all known registers.
(n-i-bz)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@12393
and use read_Type routines instead as they work rather better on strict
aligned (or semi-strict a la ARM) machines. Fixes#282527.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@12083
contains a bunch of fields which are used as a very simple state
machine that observes mmap calls and decides when to read debuginfo
for the associated file. This change moves these fields into their
own structure, struct _DebugInfoFSM, for cleanness, so as to make it
clear they have a common purpose.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@12041
info created by LLVM 2.9 work properly. As per long discussion in
#272189, this isn't actually possible -- LLVM 2.9 creates bogus line
number info, and the bogusness can't be worked around at the Valgrind
end.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@11738
knows how to unwind. This is important when unwinding Thumb code
the CFA is often stated as being at some offset from r7.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@11377
code_alignment_factor, thereby assuming it is 1. This happens to be
OK on amd64-linux and s390x-linux because it really is 1, but on
arm-linux it is 2, and hence the boundaries between code-unwind areas
are simply wrong after any of DW_CFA_advance_loc{,1,2,4} are
processed. This patch provides the obvious fix.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@11376
but with support for VLIW architectures with multiple opcodes per
instruction removed. Fixes#233595.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@11106
ARM which were originally in the loop but inadvertantly got lifted out
during recent merging. This appears to make stack unwinding work
again on ARM-Linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@10988
too. This is a first step towards making not be completely
x86/amd64-linux specific, and so replaces some x86/amd64-specific
stuff with more general constructions:
* structure 'DiCfSI', into which the info is summarised, has been
made target-specific (ugh), since the sets of registers to be
unwound differ on different targets.
* enum CfiReg and the CFIC_ constants have been expanded
accordingly, to handle both arm and x86/amd64 registers.
The abbreviation "IA" (Intel Architecture) has been used in a
few places where the x86 and amd64 definitions are shared.
* the CFI reader/summariser in readdwarf.c has been expanded &
generalised appropriately.
* the DiCfSI evaluator in debuginfo.c, VG_(use_CFI_info), has
also been generalised appropriately.
The main change is that instead of passing around triples
of (IP, SP, BP) values, a new structure 'D3UnwindRegs' is
passed around instead. This is defined differently for IA and
ARM and succeeds in hiding at least some of the differences
where we don't care about them.
Note also, D3UnwindRegs duplicates, in purpose and structure,
structure 'RegSummary' in priv_d3basics.h. This will be tidied
up in due course.
This commit almost certainly breaks stack unwinding on amd64-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@10986
versions of gcc as shipped with Fedora 12. Specific changes include:
- Vastly increase the number of opcodes we understand how to
evaluate when processing a location expression.
- Process frame unwind data from the debug_frame ELF section as
well as the eh_frame section.
- Handle version 3 CIEs in frame unwind data.
- Handle the compact form of DW_AT_data_member_location which just
gives a constant offset from the start of it's base type instead
of a full location expression.
Based on patches from Jakub Jelinek on bugs #210479 and #210566.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@10939
In addition to that it fixes a bug in restore_ctx handling, where it
was restoring the state from the same stack level in restore_ctx
context as is current in ctx, which is wrong, the CIE likely has no
DW_CFA_remember_state at all, while the FDE could have one.
(Jakub Jelinek). This is #200029, patch in comment #2.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@10697
This commit tidies up and rationalises what could be called the
"messaging" system -- that part of V to do with presenting output to
the user. In particular it brings significant improvements to XML
output.
Changes are:
* XML and normal text output now have separate file descriptors,
which solves longstanding problems for XML consumers caused by
the XML output getting polluted by unexpected non-XML output.
* This also means that we no longer have to hardwire all manner
of output settings (verbosity, etc) when XML is requested.
* The XML output format has been revised, cleaned up, and made
more suitable for use by error detecting tools in general
(various Memcheck-specific features have been removed). XML
output is enabled for Ptrcheck and Helgrind, and Memcheck is
updated to the new format.
* One side effect is that the behaviour of VG_(message) has been
made to be consistent with printf: it no longer automatically
adds a newline at the end of the output. This means multiple
calls to it can be used to build up a single line message; or a
single call can write a multi-line message. The ==pid==
preamble is automatically inserted at each newline.
* VG_(message)(Vg_UserMsg, ..args..) now has the abbreviated form
VG_(UMSG)(..args..); ditto VG_(DMSG) for Vg_DebugMsg and
VG_(EMSG) for Vg_DebugExtraMsg. A couple of other useful
printf derivatives have been added to pub_tool_libcprint.h,
most particularly VG_(vcbprintf).
* There's a small change in the core-tool interface to do with
error handling: VG_(needs_tool_errors) has a new method
void (*before_pp_Error)(Error* err) which, if non-NULL, is
called just before void (*pp_Error)(Error* err). This is to
give tools the chance to look at errors before any part of them
is printed, so they can print any XML preamble they like.
* coregrind/m_errormgr.c has been overhauled and cleaned up, and
is a bit simpler and more commented. In particular pp_Error
and VG_(maybe_record_error) are significantly changed.
The diff is huge, but mostly very boring. Most of the changes
are of the form
- VG_(message)(Vg_UserMsg, "this is a message %d", n);
+ VG_(message)(Vg_UserMsg, "this is a message %d\n", n);
Unfortunately as a result of this, it touches a large number
of source files.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@10465
