Commit Graph

67 Commits

Author SHA1 Message Date
Omar Sandoval
50f52ec295 program: make ProgramObject() arguments mandatory keywords
It's too hard to remember whether value or address comes first, so make
sure we're always explicit which one.
2018-06-28 23:40:47 -07:00
Omar Sandoval
aeb767d43e Replace lookup_variable callback with new VariableIndex class
Similar to how we replaced the lookup_type callback with TypeIndex.
2018-06-28 21:05:21 -07:00
Omar Sandoval
02c75770ea typeindex: support disambiguating types by filename 2018-06-28 19:34:10 -07:00
Omar Sandoval
0adac0747c Bring back ElfFile
Instead of constructing DwarfFile with a dict of sections, pass in an
ElfFile.
2018-06-25 23:43:14 -07:00
Omar Sandoval
26d3880708 dwarfindex: move indexing to a method instead of __init__()
We want to allow indexing files on the fly instead of all up front, so
move the indexing to a new DwarfIndex.add() method.
2018-06-25 00:25:49 -07:00
Omar Sandoval
e2caf2bb0d dwarfindex: don't index any declarations and handle specifications
For variables which are predeclared, GCC generates a DW_TAG_variable DIE
with DW_AT_name and DW_AT_declaration as well as a DW_TAG_variable DIE
without DW_AT_name but with DW_AT_specification pointing to the
declaration DIE. We should index the latter, not the former. This has a
couple of benefits: we can skip indexing variable declaration DIEs,
which contribute a lot of duplicate hash table insertions; and, we can
always get the address of a variable from DW_AT_location of the indexed
DIE instead of having to parse the symbol table.
2018-06-23 00:23:30 -07:00
Omar Sandoval
5e2d70f09e dwarfindex: add file name to DIE index entry key
A name and tag are not always enough to uniquely identify a type or
variable. For example, "struct workspace" in the Linux kernel can refer
to one of at least three types; fs/btrfs/{lzo,zlib,zstd}.c each have
their own struct workspace type. We can, however, also differentiate
DIEs on the file they were declared in.

The naive thing to do would be to include the file name as a string in
the hash table entry. However, that means we must allocate and
canonicalize each path in the line number program header and pay an
extra cache miss plus string comparison when adding a new entry.

We can get rid of the cache miss and string comparison if we instead map
the file name to a unique identifier. The foolproof way to do this would
be to create another big hash table of file names and use the hash table
entry index as the unique identifier. However, for this, we'd still need
to allocate and canoicalize each path as well as worry about another big
hash table.

Once we observe that we can get away with "almost certainly unique"
instead of "truly unique" identifiers, the next logical step is to just
use a hash of the file name as the identifier. With a 64-bit hash and
the ~50k files in the kernel, the probability of a collision is 1 in 10
billion. Even in the extremely unlikely event that there is a collision,
it only matters if the files with colliding names also have colliding
DIEs, which brings things pretty close to the realm of impossibility.

After this change, DwarfIndex.find() returns a list of DIEs matching the
name and tag. The callers will be updated to use the list in upcoming
changes.
2018-06-21 23:13:39 -07:00
Omar Sandoval
856892375a typeindex: handle pointer to const void in from_dwarf_type()
Like pointer types and function types, qualified void types don't have a
DW_AT_type.
2018-06-15 22:36:58 -07:00
Omar Sandoval
2ee6662639 program: handle pointers to typedefs of struct/union types
ProgramObject.__dir__(), member_(), and container_of_() all check that
the relevant type is a struct or union, but they all need to allow
typedefs of structs or unions, as well.
2018-05-25 22:32:03 -07:00
Omar Sandoval
6b178c5108 program: only check struct or union in ProgramObject.member_()
We're currently also doing this check in __getattr__(), which is
unnecessary overhead in the common case. We can just check the exception
in the error case.
2018-05-25 21:56:24 -07:00
Omar Sandoval
95c0682718 corereader: implement type reads in C
Now that read_memory() was converted to C, IntType.read() and friends
are a bottleneck. Convert them to C methods of CoreReader.
2018-05-25 00:41:12 -07:00
Omar Sandoval
95bde56cb7 Implement core dump reading in C
read_memory() is one of the hottest functions in profiles of tight loops
over lists of items, and it can be done much more efficiently in C.
2018-05-24 17:55:47 -07:00
Omar Sandoval
15849f5795 type: make operand_type() a method of Type
This gets rid of the huge isinstance() chain in
TypeIndex.operand_type().
2018-05-18 23:51:20 -07:00
Omar Sandoval
e7a2ba32e8 program: show dereferenced value for ProgramObject.str()
I always forget to dereference pointers, which gets annoying.
2018-05-14 20:36:19 -07:00
Omar Sandoval
54f5124be8 type: format char arrays as strings
For, e.g., comm in struct task_struct, a string is much nicer to look
at.
2018-05-14 20:35:21 -07:00
Omar Sandoval
6d980a95a7 Add better rlcompleter
The standard library rlcompleter doesn't support expressions involving
an item lookup (e.g., x[0] or x['foo']). This is a pain for the drgn
CLI, because it's common to use prog['variable'] and want to
autocomplete it. Instead of using the standard library rlcompleter,
implement our own, cleaned up version of it with the ability to handle
expressions containing [key]. rlcompleter already allows for arbitrary
__getattr__() calls, and __getitem__() isn't any different.
2018-05-13 23:51:42 -07:00
Omar Sandoval
e327aef860 program: fix ProgramObject.__round__() with ndigits is not None
In this case, we must return a ProgramObject.
2018-05-13 00:47:44 -07:00
Omar Sandoval
1916528f5d type: make offsetof() and typeof() accept arbitrary member designators
Instead of only accepting an identifier.
2018-05-13 00:42:50 -07:00
Omar Sandoval
8f46673649 Add member designator parser
A member designator is the second argument to the C offsetof() macro.
2018-05-13 00:41:20 -07:00
Omar Sandoval
90ca412a33 typeindex: fix mypy errors in drgn.program 2018-05-12 15:34:00 -07:00
Omar Sandoval
f299bea98a type/typeindex: implement saner typing for EnumType
This is a big change that makes EnumType have a compatible integer type
member instead of copying the fields, which ends up touching a lot of
stuff but also fixing a bunch of static typing errors.
2018-05-11 23:57:52 -07:00
Omar Sandoval
35271231fe program: swap ProgramObject arguments
type, address, value makes more sense and looks more like C.
2018-05-07 22:48:00 -07:00
Omar Sandoval
15ea6e8d97 program: implement ProgramObject binary operators 2018-05-07 18:48:10 -07:00
Omar Sandoval
9560b913f3 typeindex: change Type.unqualified() to TypeIndex.operand_type()
Converting an lvalue to an operand has to do a little bit more than
remove qualifiers:

- Convert array types to pointer types
- Convert function types to pointer types
2018-05-06 21:29:59 -07:00
Omar Sandoval
987f9be6db type: add Type.is_arithmetic() and Type.is_integer()
This will be used for ProgramObject operators as a shortcut for
isinstance(type.real_type(), (ArithmeticType, BitFieldType).
2018-05-06 00:28:07 -07:00
Omar Sandoval
da83e0adb3 program: add ProgramObject tests 2018-05-03 23:26:43 -07:00
Omar Sandoval
38262f8d53 typeindex: add TypeIndex.pointer() 2018-05-03 22:34:35 -07:00
Omar Sandoval
18e5e8f9bf typeindex: move easy types into base TypeIndex class
Some types don't actually have to go through find_dwarf_type(), so they
can be handled in the common code. This allows us to add a MockTypeIndex
to the tests.
2018-05-03 20:38:00 -07:00
Omar Sandoval
3cc11f3f5e typeindex: rename usual_arithmetic_conversions() to common_real_type()
It doesn't actually do the conversion, it just finds the common type.
While we're here, don't make the types unqualied; the caller should do
that.
2018-05-02 23:36:39 -07:00
Omar Sandoval
928048be97 tests: use mocked TypeIndex for type conversion tests
Instead of a real DwarfTypeIndex.
2018-05-02 23:19:05 -07:00
Omar Sandoval
4891a6341f typeindex: add method for getting type of literal
This will be used for ProgramObject operators where one operand is not a
ProgramObject.
2018-05-02 22:56:16 -07:00
Omar Sandoval
b060c9dcf7 typeindex: make sure we can find base types
The compiler might call a base type by any variation of its name (e.g.,
long or long int). Make sure we find it no matter what it chose.
2018-05-02 22:41:26 -07:00
Omar Sandoval
812951c340 type: add Type.convert()
For converting a value to a valid value of that type (i.e., casting).
2018-05-01 20:49:49 -07:00
Omar Sandoval
68462da720 typeindex: add usual arithmetic conversions
The rules are really subtle and not completely specified, so hopefully
this covers all of the corner cases... This will be used for
ProgramObject operators.
2018-05-01 00:17:20 -07:00
Omar Sandoval
78a745588a type: add TypedefType.real_type()
Instead of open-coding it in ProgramObject.
2018-04-30 22:23:55 -07:00
Omar Sandoval
c49426359b type: add Type.unqualified()
Used to get the same type with no qualifiers.
2018-04-30 22:16:44 -07:00
Omar Sandoval
26d8e0f7e9 type: add compatible integer type to EnumType 2018-04-30 19:33:40 -07:00
Omar Sandoval
cef86a18c0 Move from_dwarf_type{,_name}() logic to a class 2018-04-30 18:11:28 -07:00
Omar Sandoval
80d3058b17 type: canonicalize basic type names
Instead of using whatever the compiler emits, make sure we are
consistent with naming.
2018-04-30 00:01:18 -07:00
Omar Sandoval
e1673f9f29 typename: use a better canonical formatting for integer types
I.e., omit "int" for "short", "long", and "long long".
2018-04-29 23:56:31 -07:00
Omar Sandoval
e38ddba9c6 type: fix pretty-print of array of structs
We need to handle indentation like we do with nested structs.
2018-04-28 22:42:01 -07:00
Omar Sandoval
1d2b90e62b Document drgn.program and drgn.type more 2018-04-12 00:51:32 -07:00
Omar Sandoval
4c9a454cd2 type{,name}: get rid of __eq__()
We should only have these for testing.
2018-04-09 22:49:39 -07:00
Omar Sandoval
f347a51399 type{,name}: make qualifiers a frozenset 2018-04-09 21:57:52 -07:00
Omar Sandoval
366a706a96 type: replace Type.format() with Type.pretty(Type.read())
Most of the implementations of format() call read(). Make pretty() do
just the formatting part of format() and take the already read value.
2018-04-09 21:02:03 -07:00
Omar Sandoval
b848134c49 Add (partial) support for function types
The only part not implemented is function type name parsing.
2018-04-08 23:04:08 -07:00
Omar Sandoval
d41137d6e2 type: format empty struct without newline
This shows up in, e.g., Linux's struct lock_class_key if compiled
without lockdep.
2018-04-07 15:36:46 -07:00
Omar Sandoval
422c39823f type: allow reading incomplete array types
Just return a zero-length array.
2018-04-06 00:27:17 -07:00
Omar Sandoval
2bd81fd5fc coredump: push ELF/DWARF-specific logic out of Coredump
Instead, take callbacks for looking up variables and reading memory.
While we're at it, get rid of TypeFactory and instead implement its
methods as functions taking a DwarfIndex.
2018-04-05 00:52:00 -07:00
Omar Sandoval
b45e930336 dwarf: rewrite drgn.dwarf in pure Python
Now that DwarfIndex is in C, the rest can be simpler Python rather than
Cython.
2018-03-26 01:51:20 -07:00