Define that addresses for memory reads wrap around after the maximum
address rather than the current unpredictable behavior. This is done by:
1. Reworking drgn_memory_reader to work with an inclusive address range
so that a segment can contain UINT64_MAX. drgn_memory_reader remains
agnostic to the maximum address and requires that address ranges do
not overflow a uint64_t.
2. Adding the overflow/wrap-around logic to
drgn_program_add_memory_segment() and drgn_program_read_memory().
3. Changing direct uses of drgn_memory_reader_reader() to
drgn_program_read_memory() now that they are no longer equivalent.
(For some platforms, a fault might be more appropriate than wrapping
around, but this is a step in the right direction.)
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Linux kernel modules usually contain ELF relocations in DWARF and ORC
sections for symbols in .init sections. Since we ignore .init sections
entirely in cache_kernel_module_sections(), these relocations end up
being based on an address of 0 (so, e.g., a function from .init.text
could be reported as having an address of 0x0). It makes a little more
sense to use the address where the .init section was before it was
freed. So, let's update the sections' sh_addr but continue ignoring them
for determining the module's address range.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Use drgn_not_found where it's more appropriate, and check explicitly
against drgn_stop instead of err->code == DRGN_ERROR_STOP.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Currently, reference objects and buffer value objects have a byte order.
However, this doesn't always make sense for a couple of reasons:
- Byte order is only meaningful for scalars. What does it mean for a
struct to be big endian? A struct doesn't have a most or least
significant byte; its scalar members do.
- The DWARF specification allows either types or variables to have a
byte order (DW_AT_endianity). The only producer I could find that uses
this is GCC for the scalar_storage_order type attribute, and it only
uses it for base types, not variables. GDB only seems to use to check
it for base types, as well.
So, remove the byte order from objects, and move it to integer, boolean,
floating-point, and pointer types. This model makes more sense, and it
means that we can get the binary representation of any object now.
The only downside is that we can no longer support a bit offset for
non-scalars, but as far as I can tell, nothing needs that.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
I was going to add an Object.available_ attribute, but that made me
realize that the naming is somewhat ambiguous, as a reference object
with an invalid address might also be considered "unavailable" by users.
Use the name "absent" instead, which is more clear: the object isn't
there at all.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Now that types are associated with their program, we don't need to pass
the program separately to drgn_program_member_info() and can replace it
with a more natural drgn_type_find_member() API that takes only the type
and member name. While we're at it, get rid of drgn_member_info and
return the drgn_type_member and bit_offset directly. This also fixes a
bug that drgn_error_member_not_found() ignores the member name length.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
There are a couple of reasons that it was the wrong choice to have a
bit_offset for value objects:
1. When we store a buffer with a bit_offset, we're storing useless
padding bits.
2. bit_offset describes a location, or in other words, part of an
address. This makes sense for references, but not for values, which
are just a bag of bytes.
Get rid of union drgn_value.bit_offset in libdrgn, make
Object.bit_offset None for value objects, and disallow passing
bit_offset to the Object() constructor when creating a value. bit_offset
can still be passed when creating an object from a buffer, but we'll
shift the bytes down as necessary to store the value with no offset.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
THREAD_SIZE is still broken and I haven't looked into the root cause
(see commit 95be142d17 ("tests: disable THREAD_SIZE test")). We don't
need it anymore anyways, so let's remove it entirely.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
If the DWARF index encounters any error while parsing, it returns an
error saying only "debug information is truncated", which makes it hard
to track down parsing errors. The kmod index parser silently swallows
errors. For both, replace the mread functions with a higher-level
binary_buffer interface that can include more information including the
location of the error. For example:
/tmp/mybinary: .debug_info+0x4: expected at least 56 bytes, have 55
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Linux v5.8 changed the module section structure, so we need to get the
section name differently.
Closes#73.
Reported-by: Serapheim Dimitropoulos <serapheim@delphix.com>
Signed-off-by: Omar Sandoval <osandov@osandov.com>
If cache_kernel_module_sections() in report_loaded_kernel_module()
fails, we continue to the next iteration without advancing to the next
kernel module. Then, we fail on that same kernel module and repeat. Make
sure that we go to the next kernel module.
Fixes: 423d2cd500 ("libdrgn: dwarf_index: rework file reporting")
Reported-by: Serapheim Dimitropoulos <serapheim@delphix.com>
Signed-off-by: Omar Sandoval <osandov@osandov.com>
We're freeing path and then using it to report an error.
This has some weird knock-on effects. Since we freed the path, the error
message contains garbage. So, PyErr_SetString() can't decode it as a
UTF-8 string. The end result is a MissingDebugInfoError with no message.
Fix it by creating the error before freeing the path.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
We use /proc/modules and /sys/module to find loaded kernel modules for
the running kernel instead of walking the module list in the core dump
as an optimization. To make it easier to test the core dump path, add an
environment variable to disable the optimization.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
The next commit will allow using the offline path for the live kernel,
so the offline naming won't make much sense. Fold the offline path into
the top-level functions, and make the live path an escape hatch. Also
add some comments and improve naming for the file and directory handles
and update the coding style.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
These were added in commit e5874ad18a ("libdrgn: use libdwfl"), but
they have never been used. Remove them.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Use *_hash_pair() for hash functions that do the full double hashing and
return a struct hash_pair and hash_*() for other hashing utility
functions. Also change some of the equality function names to be more
symmetric and improve the documentation.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
I recently hit a couple of CI failures caused by relying on transitive
includes that weren't always present. include-what-you-use is a
Clang-based tool that helps with this. It's a bit finicky and noisy, so
this adds scripts/iwyu.py to make running it more convenient (but not
reliable enough to automate it in Travis).
This cleans up all reasonable include-what-you-use warnings and
reorganizes a few header files.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Debugging information tracking is currently in two places: drgn_program
finds debugging information, and drgn_dwarf_index stores it. Both of
these responsibilities make more sense as part of drgn_debug_info, so
let's move them there. This prepares us to track extra debugging
information that isn't pertinent to indexing.
This also reworks a couple of details of loading debugging information:
- drgn_dwarf_module and drgn_dwfl_module_userdata are consolidated into
a single structure, drgn_debug_info_module.
- The first pass of DWARF indexing now happens in parallel with reading
compilation units (by using OpenMP tasks).
Signed-off-by: Omar Sandoval <osandov@osandov.com>
There are a couple of related ways that we can cause undefined behavior
when parsing a malformed DWARF or depmod index file:
1. There are several places where we increment the cursor to skip past
some data. It is undefined behavior if the result points out of
bounds of the data, even if we don't attempt to dereference it.
2. read_in_bounds() checks that ptr <= end. This pointer comparison is
only defined if ptr and end both point to elements of the same array
object or one past the last element. If ptr has gone past end, then
this comparison is likely undefined anyways.
Fix it by adding a helper to skip past data with bounds checking. Then,
all of the helpers can assume that ptr <= end and maintain that
invariant. while we're here and auditing all of the call sites, let's
clean up the API and rename it from read_foo() to the less generic
mread_foo().
Signed-off-by: Omar Sandoval <osandov@osandov.com>
I originally envisioned types as dumb descriptors. This mostly works for
C because in C, types are fairly simple. However, even then the
drgn_program_member_info() API is awkward. You should be able to look up
a member directly from a type, but we need the program for caching
purposes. This has also held me back from adding offsetof() or
has_member() APIs.
Things get even messier with C++. C++ template parameters can be objects
(e.g., template <int N>). Such parameters would best be represented by a
drgn object, which we need a drgn program for. Static members are a
similar case.
So, let's reimagine types as being owned by a program. This has a few
parts:
1. In libdrgn, simple types are now created by factory functions,
drgn_foo_type_create().
2. To handle their variable length fields, compound types, enum types,
and function types are constructed with a "builder" API.
3. Simple types are deduplicated.
4. The Python type factory functions are replaced by methods of the
Program class.
5. While we're changing the API, the parameters to pointer_type() and
array_type() are reordered to be more logical (and to allow
pointer_type() to take a default size of None for the program's
default pointer size).
6. Likewise, the type factory methods take qualifiers as a keyword
argument only.
A big part of this change is updating the tests and splitting up large
test cases into smaller ones in a few places.
Signed-off-by: Omar Sandoval <osandov@osandov.com>
Declaring a local vector or hash table and separately initializing it
with vector_init()/hash_table_init() is annoying. Add macros that can be
used as initializers.
This exposes several places where the C89 style of placing all
declarations at the beginning of a block is awkward. I adopted this
style from the Linux kernel, which uses C89 and thus requires this
style. I'm now convinced that it's usually nicer to declare variables
where they're used. So let's officially adopt the style of mixing
declarations and code (and ditch the blank line after declarations) and
update the functions touched by this change.
drgn has a couple of issues unwinding stack traces for kernel core
dumps:
1. It can't unwind the stack for the idle task (PID 0), which commonly
appears in core dumps.
2. It uses the PID in PRSTATUS, which is racy and can't actually be
trusted.
The solution for both of these is to look up the PRSTATUS note by CPU
instead of PID.
For the live kernel, drgn refuses to unwind the stack of tasks in the
"R" state. However, the "R" state is running *or runnable*, so in the
latter case, we can still unwind the stack. The solution for this is to
look at on_cpu for the task instead of the state.
drgn was originally my side project, but for awhile now it's also been
my work project. Update the copyright headers to reflect this, and add a
copyright header to various files that were missing it.
Now that we can walk page tables, we can use it in a memory reader that
reads kernel memory via the kernel page table. This means that we don't
need libkdumpfile for ELF vmcores anymore (although I'll keep the
functionality around until this code has been validated more).
I originally wanted to avoid depending on another vmcoreinfo field, but
an the next change is going to depend on swapper_pg_dir in vmcoreinfo
anyways, and it ends up being simpler to use it.
Before Linux v4.11, /proc/kcore didn't have valid physical addresses, so
it's currently not possible to read from physical memory on old kernels.
However, if we can figure out the address of the direct mapping, then we
can determine the corresponding physical addresses for the segments and
add them.
We treat core dumps with all zero p_paddrs as not having valid physical
addresses. However, it is theoretically possible for a kernel core dump
to only have one segment which legitimately has a p_paddr of 0 (e.g., if
it only has a segment for the direct mapping, although note that this
isn't currently possible on x86, as Linux on x86 reserves PFN 0 for the
BIOS [1]).
If the core dump has a VMCOREINFO note, then it is either a vmcore,
which has valid physical addresses, or it is /proc/kcore with Linux
kernel commit 23c85094fe18 ("proc/kcore: add vmcoreinfo note to
/proc/kcore") (in v4.19), so it must also have Linux kernel commit
464920104bf7 ("/proc/kcore: update physical address for kcore ram and
text") (in v4.11) (ignoring the possibility of a franken-kernel which
backported the former but not the latter). Therefore, treat core dumps
with a VMCOREINFO note as having valid physical addresses.
1: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/x86/kernel/setup.c?h=v5.6#n678
Even on 32-bit architectures, physical addresses can be 64 bits (e.g.,
x86 with PAE). It's unlikely that the vmcoreinfo note would be in such
an address, but let's always parse it as a uint64_t just to be safe.
Similarly to PAGE_OFFSET, vmemmap makes more sense as part of the Linux
kernel object finder than an internal helper.
While we're here, let's fix the definition for 5-level page tables. This
only matters for kernels with commit 77ef56e4f0fb ("x86: Enable 5-level
paging support via CONFIG_X86_5LEVEL=y") but without eedb92abb9bb
("x86/mm: Make virtual memory layout dynamic for CONFIG_X86_5LEVEL=y")
(namely, v4.14, v4.15, and v4.16); since v4.17, 5-level page table
support enables KASLR.
The internal _page_offset() helper gets the value of PAGE_OFFSET, but
the fallback when KASLR is disabled has been out of date since Linux
v4.20 and never handled 5-level page tables. Additionally, it makes more
sense as part of the Linux kernel (formerly vmcoreinfo) object finder so
that it's cleanly accessible outside of drgn internals.
For types obtained from DWARF, we determine it from the language of the
CU. For other types, it can be specified manually or fall back to the
default (C). Then, we can use the language for operations where the type
is available.
UTS_RELEASE is currently only accessible once debug info is loaded with
prog.load_debug_info(main=True). This makes it difficult to get the
release, find the appropriate vmlinux, then load the found vmlinux. We
can add vmcoreinfo_object_find as part of set_core_dump(), which makes
it possible to do the following:
prog = drgn.Program()
prog.set_core_dump(core_dump_path)
release = prog['UTS_RELEASE'].string_()
vmlinux_path = find_vmlinux(release)
prog.load_debug_info([vmlinux_path])
The only downside is that this ends up using the default definition of
char rather than what we would get from the debug info, but that
shouldn't be a big problem.
The osrelease is accessible via init_uts_ns.name.release, but we can
also get it straight out of vmcoreinfo, which will be useful for the
next change. UTS_RELEASE is the name of the macro defined in the kernel.
Commit f327552229 ("libdrgn: add strstartswith()") flipped the test
for a name entry in modinfo. This introduced a regression resulting in
kernel modules not loading at the right offset. This patch fixes the
regression.
It turns out this wasn't a problem with dwfl_addrmodule() at all; the
real problem is that .init sections are freed once the module is loaded
but we're still considering them for the address range we pass to
dwfl_report_module(). Ignore those sections entirely (by omitting them
from the section name to section index map). While we're here, let's not
bother inserting non-SHF_ALLOC sections in the map.
If we only want debugging information for vmlinux and not kernel
modules, it'd be nice to only load the former. This adds a load_main
parameter to drgn_program_load_debug_info() which specifies just that.
For now, it's only implemented for the Linux kernel. While we're here,
let's make the paths parameter optional for the Python bindings.
Currently, the interface between the DWARF index, libdwfl, and the code
which finds and reports vmlinux/kernel modules is spaghetti. The DWARF
index tracks Dwfl_Modules via their userdata. However, despite
conceptually being owned by the DWARF index, the reporting code reports
the Dwfl_Modules and sets up the userdata. These Dwfl_Modules and
drgn_dwfl_module_userdatas are messy to track and pass between the
layers.
This reworks the architecture so that the DWARF index owns the Dwfl
instance and files are reported to the DWARF index; the DWARF index
takes care of reporting to libdwfl internally. In addition to making the
interface for the reporter much cleaner, this improves a few things as a
side-effect:
- We now deduplicate on build ID in addition to path.
- We now skip searching for vmlinux and/or kernel modules if they were
already indexed.
- We now support compressed ELF files via libdwelf.
- We can now load default debug info at the same time as additional
debug info.
We're too inconsistent with how we use these for them to be useful (and
it's impossible to distinguish between a format error and some other
error from libelf/libdw/libdwfl), so let's just get rid of them and make
it all DRGN_ERROR_OTHER/Exception.
struct drgn_symbol doesn't really represent a symbol; it's just an
object which hasn't been fully initialized (see c2be52dff0 ("libdrgn:
rename object index to symbol index"), it used to be called a "partial
object"). For stack traces, we're going to have a notion of a symbol
that more closely represents an ELF symbol, so let's get rid of the
temporary struct drgn_symbol representation and just return an object
directly.
We don't need to get the DWARF index at the time we get the Dwfl handle,
so get rid of drgn_program_get_dwarf(), add drgn_program_get_dwfl(), and
create the DWARF index right before we update in a new function,
drgn_program_update_dwarf_index().
libdwfl is the elfutils "DWARF frontend library". It has high-level
functionality for looking up symbols, walking stack traces, etc. In
order to use this functionality, we need to report our debugging
information through libdwfl. For userspace programs, libdwfl has a much
better implementation than drgn for automatically finding debug
information from a core dump or PID. However, for the kernel, libdwfl
has a few issues:
- It only supports finding debug information for the running kernel, not
vmcores.
- It determines the vmlinux address range by reading /proc/kallsyms,
which is slow (~70ms on my machine).
- If separate debug information isn't available for a kernel module, it
finds it by walking /lib/modules/$(uname -r)/kernel; this is repeated
for every module.
- It doesn't find kernel modules with names containing both dashes and
underscores (e.g., aes-x86_64).
Luckily, drgn already solved all of these problems, and with some
effort, we can keep doing it ourselves and report it to libdwfl.
The conversion replaces a bunch of code for dealing with userspace core
dump notes, /proc/$pid/maps, and relocations.
