JakeHillion/drgn
1
0
Fork 0
mirror of https://github.com/JakeHillion/drgn.git synced 2024-12-23 09:43:06 +00:00
Code Issues Packages Projects Releases Wiki Activity
6ad80feb83
drgn/libdrgn/linux_kernel_helpers.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

947 lines
25 KiB
C
Raw Normal View History

Update copyright headers to Meta Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-11-21 23:59:44 +00:00
// Copyright (c) Meta Platforms, Inc. and affiliates.
Relicense drgn from GPLv3+ to LGPLv2.1+ drgn is currently licensed as GPLv3+. Part of the long term vision for drgn is that other projects can use it as a library providing programmatic interfaces for debugger functionality. A more permissive license is better suited to this goal. We decided on LGPLv2.1+ as a good balance between software freedom and permissiveness. All contributors not employed by Meta were contacted via email and consented to the license change. The only exception was the author of commit c4fbf7e5896c ("libdrgn: fix for compilation error"), who did not respond. That commit reverted a single line of code to one originally written by me in commit 640b1c011df8 ("libdrgn: embed DWARF index in DWARF info cache"). Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-11-02 00:05:16 +00:00
// SPDX-License-Identifier: LGPL-2.1-or-later
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
#include <inttypes.h>
libdrgn: fix include-what-you-use warnings It's been awhile since I've run this. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-08-26 19:32:10 +01:00
#include <stdio.h>
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
Clean up #includes with include-what-you-use 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>
2020-09-24 00:02:02 +01:00
#include "drgn.h"
libdrgn: add Linux kernel task iterator The thread API needs a way to iterate over all task_structs in the kernel. Previously, we translated the existing for_each_task helper, which supports iterating through specific PID namespaces by walking through the PID radix tree or PID hashtable. However, we don't need specific namespaces for the thread API, so we can instead use the much simpler linked lists of thread groups and threads. Signed-off-by: Kevin Svetlitski <svetlitski@fb.com>
2021-12-18 00:56:00 +00:00
#include "helpers.h"
libdrgn: add {min,max}_iconst() and rewrite min() and max() min() and max() from the Linux kernel go through the trouble of resulting in a constant expression if the arguments are constant expressions, but they can't be used outside of a function due to their use of ({ }). This means that they can't be used for, e.g., enumerators or global arrays. Let's simplify min() and max() and instead add explicit min_iconst() and max_iconst() macros that can be used everywhere that an integer constant expression is required. We can then use it in hash_table.h. While we're here, let's split these into their own header file and document them better. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-10-10 01:59:23 +01:00
#include "minmax.h"
Clean up #includes with include-what-you-use 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>
2020-09-24 00:02:02 +01:00
#include "platform.h"
libdrgn: add task_state_to_char() helper Add a helper to get the state of a task (e.g., 'R', 'S', 'D'). This will be used to make sure that a task is not running when getting a stack trace, so implement it in libdrgn.
2019-10-28 09:38:51 +00:00
#include "program.h"
Fix various lints Some functions that could be static found by -Wmissing-prototypes, some include-what-you-use warnings, some missing SPDX identifiers. These lints should be automated at some point. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-02-17 18:45:42 +00:00
#include "util.h"
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
static void end_virtual_address_translation(struct drgn_program *prog)
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
{
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
prog->in_address_translation = false;
}
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
static struct drgn_error *
begin_virtual_address_translation(struct drgn_program *prog, uint64_t pgtable,
uint64_t virt_addr)
{
struct drgn_error *err;
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
if (prog->in_address_translation) {
libdrgn: add page table walker kernel memory reader 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).
2020-05-06 08:35:29 +01:00
return drgn_error_create_fault("recursive address translation; "
"page table may be missing from core dump",
virt_addr);
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
prog->in_address_translation = true;
if (!prog->pgtable_it) {
if (!(prog->flags & DRGN_PROGRAM_IS_LINUX_KERNEL)) {
err = drgn_error_create(DRGN_ERROR_INVALID_ARGUMENT,
"virtual address translation is only available for the Linux kernel");
goto err;
}
if (!prog->has_platform) {
err = drgn_error_create(DRGN_ERROR_INVALID_ARGUMENT,
"cannot do virtual address translation without platform");
goto err;
}
if (!prog->platform.arch->linux_kernel_pgtable_iterator_next) {
err = drgn_error_format(DRGN_ERROR_INVALID_ARGUMENT,
"virtual address translation is not implemented for %s architecture",
prog->platform.arch->name);
goto err;
}
err = prog->platform.arch->linux_kernel_pgtable_iterator_create(prog,
&prog->pgtable_it);
if (err) {
prog->pgtable_it = NULL;
goto err;
}
}
prog->pgtable_it->pgtable = pgtable;
prog->pgtable_it->virt_addr = virt_addr;
prog->platform.arch->linux_kernel_pgtable_iterator_init(prog, prog->pgtable_it);
return NULL;
libdrgn: add page table walker kernel memory reader 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).
2020-05-06 08:35:29 +01:00
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
err:
end_virtual_address_translation(prog);
return err;
}
libdrgn: linux_kernel: make virt_to_phys() and phys_to_virt() generic On x86-64, the difference between virtual addresses in the direct map and the corresponding physical addresses is called PAGE_OFFSET, so we exposed that via an architecture callback and the Linux kernel object finder. However, this doesn't translate to other architectures. Namely, on AArch64, the difference is PAGE_OFFSET - PHYS_OFFSET, and both PAGE_OFFSET and PHYS_OFFSET have varied over time and between configurations. We can remove the architecture callback and avoid version-specific logic by letting the page table tell us the offset. We just need an address in the direct map, which is easy to find since this includes kmalloc and memblock allocations. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-02 21:14:18 +01:00
struct drgn_error *linux_helper_direct_mapping_offset(struct drgn_program *prog,
uint64_t *ret)
{
struct drgn_error *err;
if (prog->direct_mapping_offset_cached) {
*ret = prog->direct_mapping_offset;
return NULL;
}
// The direct mapping offset can vary depending on architecture, kernel
// version, configuration options, and KASLR. Rather than dealing with
// all of that, get a virtual address in the direct mapping and
// translate it to a physical address via the page table. The difference
// is the offset.
//
// The virtual address we pick doesn't matter as long as:
//
// 1. It is in the direct mapping on all configurations of all supported
// kernel versions on all architectures.
// 2. That is unlikely to change in the future.
//
// This is our current arbitrary choice.
static const char direct_mapping_variable[] = "saved_command_line";
struct drgn_object tmp;
drgn_object_init(&tmp, prog);
err = drgn_program_find_object(prog, direct_mapping_variable, NULL,
DRGN_FIND_OBJECT_VARIABLE, &tmp);
uint64_t virt_addr;
if (!err) {
err = drgn_object_read_unsigned(&tmp, &virt_addr);
} else if (err->code == DRGN_ERROR_LOOKUP) {
// Avoid a confusing error message with our arbitrary variable
// name.
drgn_error_destroy(err);
err = drgn_error_create(DRGN_ERROR_OTHER,
"could not find variable in direct mapping");
}
drgn_object_deinit(&tmp);
if (err)
return err;
err = begin_virtual_address_translation(prog,
prog->vmcoreinfo.swapper_pg_dir,
virt_addr);
if (err)
return err;
uint64_t start_virt_addr, start_phys_addr;
err = prog->platform.arch->linux_kernel_pgtable_iterator_next(prog,
prog->pgtable_it,
&start_virt_addr,
&start_phys_addr);
if (err)
goto out;
if (start_phys_addr == UINT64_MAX) {
err = drgn_error_create(DRGN_ERROR_OTHER,
"could not determine direct mapping offset");
goto out;
}
prog->direct_mapping_offset = start_virt_addr - start_phys_addr;
prog->direct_mapping_offset_cached = true;
*ret = prog->direct_mapping_offset;
err = NULL;
out:
end_virtual_address_translation(prog);
return err;
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
struct drgn_error *linux_helper_read_vm(struct drgn_program *prog,
uint64_t pgtable, uint64_t virt_addr,
void *buf, size_t count)
{
struct drgn_error *err;
err = begin_virtual_address_translation(prog, pgtable, virt_addr);
if (err)
return err;
if (!count) {
err = NULL;
goto out;
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
struct pgtable_iterator *it = prog->pgtable_it;
pgtable_iterator_next_fn *next =
prog->platform.arch->linux_kernel_pgtable_iterator_next;
uint64_t read_addr = 0;
size_t read_size = 0;
do {
uint64_t start_virt_addr, start_phys_addr;
err = next(prog, it, &start_virt_addr, &start_phys_addr);
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
if (err)
break;
if (start_phys_addr == UINT64_MAX) {
err = drgn_error_create_fault("address is not mapped",
virt_addr);
break;
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
libdrgn: linux_kernel: fix read_vm() coalescing comparison linux_helper_read_vm() has logic to merge adjacent physical address ranges returned by the page table iterator. However, the check for whether the ranges are adjacent is incorrect. Fix it. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:42:14 +01:00
uint64_t phys_addr =
start_phys_addr + (virt_addr - start_virt_addr);
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
size_t n = min(it->virt_addr - virt_addr, (uint64_t)count);
libdrgn: linux_kernel: fix read_vm() coalescing comparison linux_helper_read_vm() has logic to merge adjacent physical address ranges returned by the page table iterator. However, the check for whether the ranges are adjacent is incorrect. Fix it. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:42:14 +01:00
if (read_size && phys_addr == read_addr + read_size) {
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
read_size += n;
} else {
if (read_size) {
err = drgn_program_read_memory(prog, buf,
read_addr,
read_size, true);
if (err)
break;
buf = (char *)buf + read_size;
}
libdrgn: linux_kernel: fix read_vm() coalescing comparison linux_helper_read_vm() has logic to merge adjacent physical address ranges returned by the page table iterator. However, the check for whether the ranges are adjacent is incorrect. Fix it. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:42:14 +01:00
read_addr = phys_addr;
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
read_size = n;
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
virt_addr = it->virt_addr;
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
count -= n;
} while (count);
if (!err) {
err = drgn_program_read_memory(prog, buf, read_addr, read_size,
true);
}
libdrgn: refactor page table iterators AArch64 will need different sizes of page table iterators depending on the page size and virtual address size. Rather than the static pgtable_iterator_arch_size, allow architectures to define callbacks for allocating and freeing a page table iterator. Also remove the generic page table iterator wrapper and just pass that information to the iterator function. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-07-08 23:34:29 +01:00
out:
end_virtual_address_translation(prog);
helpers: add access_process_vm() and access_remote_vm() Now that we can walk page tables, we can finally read memory from userspace tasks. Closes #53.
2020-05-06 02:21:36 +01:00
return err;
}
libdrgn: linux: translate per_cpu_ptr() helper to C The next change will add a C helper that needs per_cpu_ptr(). Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:39:50 +00:00
struct drgn_error *linux_helper_per_cpu_ptr(struct drgn_object *res,
const struct drgn_object *ptr,
uint64_t cpu)
{
struct drgn_error *err;
struct drgn_program *prog = drgn_object_program(ptr);
struct drgn_object tmp;
drgn_object_init(&tmp, prog);
err = drgn_program_find_object(prog, "__per_cpu_offset", NULL,
DRGN_FIND_OBJECT_ANY, &tmp);
libdrgn: linux: make per_cpu_ptr() support !SMP kernels Kernels built without multiprocessing support don't have __per_cpu_offset; instead, per_cpu_ptr() is a no-op. Make the helper do the same and update the test case to work on !SMP as well. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:51:15 +00:00
if (!err) {
err = drgn_object_subscript(&tmp, &tmp, cpu);
if (err)
goto out;
union drgn_value per_cpu_offset;
err = drgn_object_read_integer(&tmp, &per_cpu_offset);
if (err)
goto out;
libdrgn: linux: translate per_cpu_ptr() helper to C The next change will add a C helper that needs per_cpu_ptr(). Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:39:50 +00:00
libdrgn: linux: make per_cpu_ptr() support !SMP kernels Kernels built without multiprocessing support don't have __per_cpu_offset; instead, per_cpu_ptr() is a no-op. Make the helper do the same and update the test case to work on !SMP as well. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:51:15 +00:00
uint64_t ptr_value;
err = drgn_object_read_unsigned(ptr, &ptr_value);
if (err)
goto out;
libdrgn: linux: translate per_cpu_ptr() helper to C The next change will add a C helper that needs per_cpu_ptr(). Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:39:50 +00:00
libdrgn: linux: make per_cpu_ptr() support !SMP kernels Kernels built without multiprocessing support don't have __per_cpu_offset; instead, per_cpu_ptr() is a no-op. Make the helper do the same and update the test case to work on !SMP as well. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:51:15 +00:00
err = drgn_object_set_unsigned(res,
drgn_object_qualified_type(ptr),
ptr_value + per_cpu_offset.uvalue,
0);
} else if (err->code == DRGN_ERROR_LOOKUP) {
drgn_error_destroy(err);
err = drgn_object_copy(res, ptr);
}
libdrgn: linux: translate per_cpu_ptr() helper to C The next change will add a C helper that needs per_cpu_ptr(). Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:39:50 +00:00
out:
drgn_object_deinit(&tmp);
return err;
}
libdrgn: linux: replace idle_thread() with idle_task() I missed that the kernel has an idle_task() function which uses cpu_rq()->idle instead of idle_threads; the latter is technically architecture-specific. So, replace idle_thread() with idle_task(), which is architecture-independent and more consistent with the kernel. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:03:25 +00:00
struct drgn_error *linux_helper_idle_task(struct drgn_object *res, uint64_t cpu)
libdrgn: linux: add idle_thread() helper PR #129 will need to get the idle thread for a CPU when the idle thread crashed. Add a helper for this. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:40:57 +00:00
{
struct drgn_error *err;
struct drgn_program *prog = drgn_object_program(res);
struct drgn_object tmp;
drgn_object_init(&tmp, prog);
libdrgn: linux: replace idle_thread() with idle_task() I missed that the kernel has an idle_task() function which uses cpu_rq()->idle instead of idle_threads; the latter is technically architecture-specific. So, replace idle_thread() with idle_task(), which is architecture-independent and more consistent with the kernel. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:03:25 +00:00
err = drgn_program_find_object(prog, "runqueues", NULL,
libdrgn: linux: add idle_thread() helper PR #129 will need to get the idle thread for a CPU when the idle thread crashed. Add a helper for this. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:40:57 +00:00
DRGN_FIND_OBJECT_ANY, &tmp);
libdrgn: linux: replace idle_thread() with idle_task() I missed that the kernel has an idle_task() function which uses cpu_rq()->idle instead of idle_threads; the latter is technically architecture-specific. So, replace idle_thread() with idle_task(), which is architecture-independent and more consistent with the kernel. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-22 00:03:25 +00:00
if (err)
goto out;
err = drgn_object_address_of(&tmp, &tmp);
if (err)
goto out;
err = linux_helper_per_cpu_ptr(&tmp, &tmp, cpu);
if (err)
goto out;
err = drgn_object_member_dereference(res, &tmp, "idle");
libdrgn: linux: add idle_thread() helper PR #129 will need to get the idle thread for a CPU when the idle thread crashed. Add a helper for this. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-21 22:40:57 +00:00
out:
drgn_object_deinit(&tmp);
return err;
}
Fix vmcore stack traces on Linux < 4.9 or >= 5.16 and add drgn.helpers.linux.task_cpu() task->cpu was moved to task->thread_info.cpu in Linux 5.16, which causes drgn_get_initial_registers() to think that the kernel is !SMP and use CPU 0 instead, producing incorrect stack traces. This has also always been wrong for Linux < 4.9 and on architectures that don't enable CONFIG_THREAD_INFO_IN_TASK; in those cases, it should be ((struct thread_info *)task->stack)->cpu. Fix it by factoring out a new task_cpu() helper that handles all of the above cases. Also add a test case for task_cpu() in case this changes again. Fixes: eea542254600 ("libdrgn: make Linux kernel stack unwinding more robust") Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-10-04 00:08:53 +01:00
struct drgn_error *linux_helper_task_cpu(const struct drgn_object *task,
uint64_t *ret)
{
struct drgn_error *err;
struct drgn_object tmp;
drgn_object_init(&tmp, drgn_object_program(task));
// If CONFIG_THREAD_INFO_IN_TASK=y and since Linux kernel commit
// bcf9033e5449 ("sched: move CPU field back into thread_info if
// THREAD_INFO_IN_TASK=y") (in v5.16), the CPU is task->thread_info.cpu.
//
// If CONFIG_THREAD_INFO_IN_TASK=y but before that commit, the cpu is
// task->cpu.
//
// If CONFIG_THREAD_INFO_IN_TASK=n or before Linux kernel commit
// c65eacbe290b ("sched/core: Allow putting thread_info into
// task_struct") (in v4.9), the CPU is
// ((struct thread_info *)task->stack)->cpu.
//
// If none of those exist, then the kernel must be !SMP.
err = drgn_object_member_dereference(&tmp, task, "thread_info");
if (!err) {
err = drgn_object_member(&tmp, &tmp, "cpu");
if (err && err->code == DRGN_ERROR_LOOKUP) {
drgn_error_destroy(err);
err = drgn_object_member_dereference(&tmp, task, "cpu");
}
} else if (err->code == DRGN_ERROR_LOOKUP) {
// CONFIG_THREAD_INFO_IN_TASK=n
drgn_error_destroy(err);
err = drgn_object_member_dereference(&tmp, task, "stack");
if (err)
goto out;
struct drgn_qualified_type thread_info_type;
err = drgn_program_find_type(drgn_object_program(task),
"struct thread_info *", NULL,
&thread_info_type);
if (err)
goto out;
err = drgn_object_cast(&tmp, thread_info_type, &tmp);
if (err)
goto out;
err = drgn_object_member_dereference(&tmp, &tmp, "cpu");
}
if (!err) {
union drgn_value value;
err = drgn_object_read_integer(&tmp, &value);
if (!err)
*ret = value.uvalue;
} else if (err->code == DRGN_ERROR_LOOKUP) {
// CONFIG_SMP=n
drgn_error_destroy(err);
*ret = 0;
err = NULL;
}
out:
drgn_object_deinit(&tmp);
return err;
}
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
struct drgn_error *
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
linux_helper_xa_load(struct drgn_object *res,
const struct drgn_object *xa, uint64_t index)
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
{
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
struct drgn_error *err;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
struct drgn_object entry, node, tmp;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
struct drgn_qualified_type node_type;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
uint64_t internal_flag, node_min;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
drgn_object_init(&entry, drgn_object_program(res));
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
drgn_object_init(&node, drgn_object_program(res));
drgn_object_init(&tmp, drgn_object_program(res));
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
// See xa_for_each() in drgn/helpers/linux/xarray.py for a description
// of the cases we have to handle.
// entry = xa->xa_head
err = drgn_object_member_dereference(&entry, xa, "xa_head");
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (!err) {
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
err = drgn_object_read(&entry, &entry);
if (err)
goto out;
// node_type = struct xa_node *
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
err = drgn_program_find_type(drgn_object_program(res),
"struct xa_node *", NULL,
&node_type);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
internal_flag = 2;
node_min = 4097;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
} else if (err->code == DRGN_ERROR_LOOKUP) {
drgn_error_destroy(err);
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
// entry = (void *)xa->rnode
err = drgn_object_member_dereference(&entry, xa, "rnode");
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
// node_type = typeof(xa->rnode)
node_type = drgn_object_qualified_type(&entry);
struct drgn_qualified_type voidp_type;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
err = drgn_program_find_type(drgn_object_program(res), "void *",
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
NULL, &voidp_type);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
err = drgn_object_cast(&entry, voidp_type, &entry);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
internal_flag = 1;
node_min = 0;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
} else {
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
}
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
// xa_is_node() or radix_tree_is_internal_node()
#define is_node(entry_value) \
(((entry_value) & 3) == internal_flag && (entry_value) >= node_min)
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_type_member *member;
uint64_t member_bit_offset;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
err = drgn_type_find_member(drgn_type_type(node_type.type).type,
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
"slots", &member, &member_bit_offset);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_qualified_type member_type;
libdrgn: get drgn_type_member.bit_field_size through drgn_member_type() Getting the bit field size of a member will soon require evaluating the lazy type, so return it from drgn_member_type() instead of accessing it directly. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-12-18 22:11:29 +00:00
err = drgn_member_type(member, &member_type, NULL);
Rewrite linux helper iterators in C In preparation for introducing an API to represent threads, the linux helper iterators, radix_tree_for_each, idr_for_each, for_each_pid, and for_each_task have been rewritten in C. This will allow them to be accessed from libdrgn, which will be necessary for the threads API. Signed-off-by: Kevin Svetlitski <svetlitski@fb.com>
2021-11-19 00:46:59 +00:00
if (err)
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
goto out;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
if (drgn_type_kind(member_type.type) != DRGN_TYPE_ARRAY) {
err = drgn_error_create(DRGN_ERROR_TYPE,
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
"struct xa_node slots member is not an array");
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
goto out;
}
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
uint64_t XA_CHUNK_MASK = drgn_type_length(member_type.type) - 1;
uint64_t sizeof_slots;
if (node_min == 0) { // !xarray
err = drgn_type_sizeof(member_type.type, &sizeof_slots);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
}
uint64_t entry_value;
err = drgn_object_read_unsigned(&entry, &entry_value);
if (err)
goto out;
if (is_node(entry_value)) {
// node = xa_to_node(entry)
// or
// node = entry_to_node(entry)
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = drgn_object_set_unsigned(&node, node_type,
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
entry_value - internal_flag, 0);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
// node_shift = node->shift
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = drgn_object_member_dereference(&tmp, &node, "shift");
if (err)
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
union drgn_value node_shift;
err = drgn_object_read_integer(&tmp, &node_shift);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
uint64_t offset;
if (node_shift.uvalue >= 64) // Avoid undefined behavior.
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
offset = 0;
else
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
offset = index >> node_shift.uvalue;
if (offset > XA_CHUNK_MASK)
goto null;
for (;;) {
// entry = node->slots[offset]
err = drgn_object_member_dereference(&tmp, &node,
"slots");
if (err)
goto out;
err = drgn_object_subscript(&entry, &tmp, offset);
if (err)
goto out;
err = drgn_object_read(&entry, &entry);
if (err)
goto out;
err = drgn_object_read_unsigned(&entry, &entry_value);
if (err)
goto out;
if ((entry_value & 3) == internal_flag) {
if (node_min != 0 && // xarray
entry_value < 256) { // xa_is_sibling()
// entry = node->slots[xa_to_sibling(entry)]
err = drgn_object_subscript(&entry,
&tmp,
entry_value >> 2);
if (err)
goto out;
err = drgn_object_read(&entry, &entry);
if (err)
goto out;
err = drgn_object_read_unsigned(&entry,
&entry_value);
if (err)
goto out;
} else if (node_min == 0 && // !xarray
tmp.address <= entry_value &&
entry_value < tmp.address + sizeof_slots) { // is_sibling_entry()
// entry = *(void **)entry_to_node(entry)
struct drgn_qualified_type voidpp_type;
err = drgn_program_find_type(drgn_object_program(res),
"void **",
NULL,
&voidpp_type);
if (err)
goto out;
err = drgn_object_set_unsigned(&entry,
voidpp_type,
entry_value - 1,
0);
if (err)
goto out;
err = drgn_object_dereference(&entry,
&entry);
if (err)
goto out;
err = drgn_object_read(&entry, &entry);
if (err)
goto out;
err = drgn_object_read_unsigned(&entry,
&entry_value);
if (err)
goto out;
}
}
if (node_shift.uvalue == 0 || !is_node(entry_value))
break;
// node = xa_to_node(entry)
// or
// node = entry_to_node(entry)
err = drgn_object_set_unsigned(&node, node_type,
entry_value - internal_flag,
0);
if (err)
goto out;
// node_shift = node->shift
err = drgn_object_member_dereference(&tmp, &node,
"shift");
if (err)
goto out;
err = drgn_object_read_integer(&tmp, &node_shift);
if (err)
goto out;
if (node_shift.uvalue >= 64) // Avoid undefined behavior.
offset = 0;
else
offset = (index >> node_shift.uvalue) & XA_CHUNK_MASK;
}
} else if (index) {
goto null;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
}
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
err = drgn_object_copy(res, &entry);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
out:
drgn_object_deinit(&tmp);
drgn_object_deinit(&node);
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
drgn_object_deinit(&entry);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
return err;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
null:
err = drgn_object_set_unsigned(res, drgn_object_qualified_type(&entry),
0, 0);
goto out;
#undef is_node
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
}
struct drgn_error *linux_helper_idr_find(struct drgn_object *res,
const struct drgn_object *idr,
uint64_t id)
{
struct drgn_error *err;
struct drgn_object tmp;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
drgn_object_init(&tmp, drgn_object_program(res));
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
/* id -= idr->idr_base */
err = drgn_object_member_dereference(&tmp, idr, "idr_base");
if (!err) {
union drgn_value idr_base;
err = drgn_object_read_integer(&tmp, &idr_base);
if (err)
goto out;
id -= idr_base.uvalue;
} else if (err->code == DRGN_ERROR_LOOKUP) {
/* idr_base was added in v4.16. */
drgn_error_destroy(err);
} else {
goto out;
}
/* radix_tree_lookup(&idr->idr_rt, id) */
err = drgn_object_member_dereference(&tmp, idr, "idr_rt");
if (err)
goto out;
err = drgn_object_address_of(&tmp, &tmp);
if (err)
goto out;
drgn.helpers.linux: add proper XArray helpers Commit 89eb868e9519 ("helpers: make find_task() work on recent kernels") made radix_tree_lookup() and radix_tree_for_each() work for basic XArrays. However, it doesn't handle a couple of more advanced features: multi-index entries (which old radix trees actually also supported) and zero entries. It has also been really confusing to explain to people unfamiliar with the radix tree -> XArray transition that they should use helpers named radix_tree for a structure named xarray. So, let's finally add xa_load(), xa_for_each(), and some additional auxiliary helpers. The non-recursive xa_for_each() implementation is based on Kevin Svetlitski's C implementation from commit 2b47583c732a ("Rewrite linux helper iterators in C"). radix_tree_lookup() and radix_tree_for_each() share the implementation with xa_load() and xa_for_each(), respectively, so they are mostly interchangeable. Fixes: #61 Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-12-14 01:10:22 +00:00
err = linux_helper_xa_load(res, &tmp, id);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
out:
drgn_object_deinit(&tmp);
return err;
}
/*
* Before Linux kernel commit 95846ecf9dac ("pid: replace pid bitmap
* implementation with IDR API") (in v4.15), (struct pid_namespace).idr does not
* exist, so we have to search pid_hash. We could implement pid_hashfn() and
* only search that bucket, but it's different for 32-bit and 64-bit systems,
* and it has changed at least once, in v4.7. Searching the whole hash table is
* slower but foolproof.
*/
static struct drgn_error *
find_pid_in_pid_hash(struct drgn_object *res, const struct drgn_object *ns,
const struct drgn_object *pid_hash, uint64_t pid)
{
struct drgn_error *err;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_qualified_type pidp_type;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
err = drgn_program_find_type(drgn_object_program(res), "struct pid *",
NULL, &pidp_type);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
return err;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
struct drgn_qualified_type upid_type;
err = drgn_program_find_type(drgn_object_program(res), "struct upid",
NULL, &upid_type);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
return err;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_type_member *pid_chain_member;
uint64_t pid_chain_bit_offset;
err = drgn_type_find_member(upid_type.type, "pid_chain",
&pid_chain_member, &pid_chain_bit_offset);
if (err)
return err;
struct drgn_type_member *nr_member;
uint64_t nr_bit_offset;
err = drgn_type_find_member(upid_type.type, "nr", &nr_member,
&nr_bit_offset);
if (err)
return err;
struct drgn_qualified_type nr_type;
libdrgn: get drgn_type_member.bit_field_size through drgn_member_type() Getting the bit field size of a member will soon require evaluating the lazy type, so return it from drgn_member_type() instead of accessing it directly. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-12-18 22:11:29 +00:00
err = drgn_member_type(nr_member, &nr_type, NULL);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
return err;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_type_member *ns_member;
uint64_t ns_bit_offset;
err = drgn_type_find_member(upid_type.type, "ns", &ns_member,
&ns_bit_offset);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
return err;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
struct drgn_qualified_type ns_type;
libdrgn: get drgn_type_member.bit_field_size through drgn_member_type() Getting the bit field size of a member will soon require evaluating the lazy type, so return it from drgn_member_type() instead of accessing it directly. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-12-18 22:11:29 +00:00
err = drgn_member_type(ns_member, &ns_type, NULL);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
return err;
Revert "Rewrite linux helper iterators in C" This reverts commit 2b47583c732a1ad63fe544a08feddd2ea96ef1a5. After Kevin had completed this, we realized that there is a simpler method for iterating through tasks from libdrgn, which the next commit will implement. Revert the translation, but keep the improved tests.helpers.linux.test_pid.TestPid.test_for_each_task. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-01-10 20:13:04 +00:00
struct drgn_object node, tmp;
drgn_object_init(&node, drgn_object_program(res));
drgn_object_init(&tmp, drgn_object_program(res));
err = drgn_object_read(&tmp, ns);
if (err)
goto out;
uint64_t ns_addr;
err = drgn_object_read_unsigned(&tmp, &ns_addr);
if (err)
goto out;
union drgn_value ns_level;
err = drgn_object_member_dereference(&tmp, &tmp, "level");
if (err)
goto out;
err = drgn_object_read_integer(&tmp, &ns_level);
if (err)
goto out;
/* i = 1 << pidhash_shift */
err = drgn_program_find_object(drgn_object_program(res),
"pidhash_shift", NULL,
DRGN_FIND_OBJECT_ANY, &tmp);
if (err)
goto out;
union drgn_value pidhash_shift;
err = drgn_object_read_integer(&tmp, &pidhash_shift);
if (err)
goto out;
uint64_t i;
if (pidhash_shift.uvalue >= 64)
i = 0;
else
i = UINT64_C(1) << pidhash_shift.uvalue;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
while (i--) {
/* for (node = pid_hash[i].first; node; node = node->next) */
err = drgn_object_subscript(&node, pid_hash, i);
if (err)
goto out;
err = drgn_object_member(&node, &node, "first");
if (err)
goto out;
for (;;) {
uint64_t addr, tmp_addr;
union drgn_value node_nr;
uint64_t node_ns;
err = drgn_object_read(&node, &node);
if (err)
goto out;
err = drgn_object_read_unsigned(&node, &addr);
if (err)
goto out;
if (!addr)
break;
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
addr -= pid_chain_bit_offset / 8;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
/* tmp = container_of(node, struct upid, pid_chain)->nr */
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
tmp_addr = addr + nr_bit_offset / 8;
err = drgn_object_set_reference(&tmp, nr_type, tmp_addr,
Track byte order in scalar types instead of objects 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>
2021-02-18 00:13:23 +00:00
0, 0);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
goto out;
err = drgn_object_read_integer(&tmp, &node_nr);
if (err)
goto out;
if (node_nr.uvalue != pid)
goto next;
/* tmp = container_of(node, struct upid, pid_chain)->ns */
libdrgn: replace drgn_program_member_info() with drgn_type_find_member() 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>
2020-12-15 01:58:44 +00:00
tmp_addr = addr + ns_bit_offset / 8;
err = drgn_object_set_reference(&tmp, ns_type, tmp_addr,
Track byte order in scalar types instead of objects 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>
2021-02-18 00:13:23 +00:00
0, 0);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (err)
goto out;
err = drgn_object_read_unsigned(&tmp, &node_ns);
if (err)
goto out;
if (node_ns != ns_addr)
goto next;
libdrgn: use exact buffer sizes when formatting decimal numbers We have a few places where we format a decimal number with sprintf() or snprintf() to a buffer with an arbitrary size. Instead of this arbitrary size, let's add a macro to get the exact number of characters required to format a decimal number, use it in all of these places, and make all of these places use snprintf() just to be safe. This is more verbose but self-documenting. The max_decimal_length() macro is inspired by https://stackoverflow.com/a/13546502/1811295 with some improvements. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-02-11 10:59:41 +00:00
#define FORMAT "numbers[%" PRIu64 "].pid_chain"
char member[sizeof(FORMAT)
- sizeof("%" PRIu64)
+ max_decimal_length(uint64_t)
+ 1];
snprintf(member, sizeof(member), FORMAT,
ns_level.uvalue);
#undef FORMAT
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = drgn_object_container_of(res, &node,
drgn_type_type(pidp_type.type),
member);
goto out;
next:
err = drgn_object_member_dereference(&node, &node, "next");
if (err)
goto out;
}
}
err = drgn_object_set_unsigned(res, pidp_type, 0, 0);
out:
drgn_object_deinit(&tmp);
drgn_object_deinit(&node);
return err;
}
struct drgn_error *linux_helper_find_pid(struct drgn_object *res,
const struct drgn_object *ns,
uint64_t pid)
{
struct drgn_error *err;
struct drgn_object tmp;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
drgn_object_init(&tmp, drgn_object_program(res));
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
/* (struct pid *)idr_find(&ns->idr, pid) */
err = drgn_object_member_dereference(&tmp, ns, "idr");
if (!err) {
struct drgn_qualified_type qualified_type;
err = drgn_object_address_of(&tmp, &tmp);
if (err)
goto out;
err = linux_helper_idr_find(&tmp, &tmp, pid);
if (err)
goto out;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
err = drgn_program_find_type(drgn_object_program(res),
"struct pid *", NULL,
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
&qualified_type);
if (err)
goto out;
err = drgn_object_cast(res, qualified_type, &tmp);
} else if (err->code == DRGN_ERROR_LOOKUP) {
drgn_error_destroy(err);
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
err = drgn_program_find_object(drgn_object_program(res),
"pid_hash", NULL,
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
DRGN_FIND_OBJECT_ANY, &tmp);
if (err)
goto out;
err = find_pid_in_pid_hash(res, ns, &tmp, pid);
}
out:
drgn_object_deinit(&tmp);
return err;
}
struct drgn_error *linux_helper_pid_task(struct drgn_object *res,
const struct drgn_object *pid,
uint64_t pid_type)
{
struct drgn_error *err;
struct drgn_qualified_type task_structp_type;
struct drgn_qualified_type task_struct_type;
bool truthy;
struct drgn_object first;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
drgn_object_init(&first, drgn_object_program(res));
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
err = drgn_program_find_type(drgn_object_program(res),
"struct task_struct *", NULL,
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
&task_structp_type);
if (err)
goto out;
task_struct_type = drgn_type_type(task_structp_type.type);
err = drgn_object_bool(pid, &truthy);
libdrgn: helpers: add missing error check in linux_helper_pid_task() Found with clang-static-analyzer. Reported-by: Kevin Svetlitski <svetlitski@fb.com> Signed-off-by: Omar Sandoval <osandov@osandov.com>
2021-12-08 21:37:01 +00:00
if (err)
goto out;
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
if (!truthy)
goto null;
/* first = &pid->tasks[pid_type].first */
err = drgn_object_member_dereference(&first, pid, "tasks");
if (err)
goto out;
err = drgn_object_subscript(&first, &first, pid_type);
if (err)
goto out;
err = drgn_object_member(&first, &first, "first");
if (err)
goto out;
err = drgn_object_bool(&first, &truthy);
if (err)
goto out;
if (!truthy)
goto null;
/* container_of(first, struct task_struct, pid_links[pid_type]) */
libdrgn: use exact buffer sizes when formatting decimal numbers We have a few places where we format a decimal number with sprintf() or snprintf() to a buffer with an arbitrary size. Instead of this arbitrary size, let's add a macro to get the exact number of characters required to format a decimal number, use it in all of these places, and make all of these places use snprintf() just to be safe. This is more verbose but self-documenting. The max_decimal_length() macro is inspired by https://stackoverflow.com/a/13546502/1811295 with some improvements. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-02-11 10:59:41 +00:00
#define PID_LINKS_FORMAT "pid_links[%" PRIu64 "]"
#define PIDS_NODE_FORMAT "pids[%" PRIu64 "].node"
char member[max_iconst(sizeof(PID_LINKS_FORMAT),
sizeof(PIDS_NODE_FORMAT))
- sizeof("%" PRIu64)
+ max_decimal_length(uint64_t)
+ 1];
snprintf(member, sizeof(member), PID_LINKS_FORMAT, pid_type);
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = drgn_object_container_of(res, &first, task_struct_type, member);
if (err && err->code == DRGN_ERROR_LOOKUP) {
drgn_error_destroy(err);
/* container_of(first, struct task_struct, pids[pid_type].node) */
libdrgn: use exact buffer sizes when formatting decimal numbers We have a few places where we format a decimal number with sprintf() or snprintf() to a buffer with an arbitrary size. Instead of this arbitrary size, let's add a macro to get the exact number of characters required to format a decimal number, use it in all of these places, and make all of these places use snprintf() just to be safe. This is more verbose but self-documenting. The max_decimal_length() macro is inspired by https://stackoverflow.com/a/13546502/1811295 with some improvements. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2022-02-11 10:59:41 +00:00
snprintf(member, sizeof(member), PIDS_NODE_FORMAT, pid_type);
#undef PID_LINKS_FORMAT
#undef PIDS_NODE_FORMAT
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = drgn_object_container_of(res, &first, task_struct_type,
member);
}
out:
drgn_object_deinit(&first);
return err;
null:
err = drgn_object_set_unsigned(res, task_structp_type, 0, 0);
goto out;
}
struct drgn_error *linux_helper_find_task(struct drgn_object *res,
const struct drgn_object *ns,
uint64_t pid)
{
struct drgn_error *err;
struct drgn_object pid_obj;
struct drgn_object pid_type_obj;
union drgn_value pid_type;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
drgn_object_init(&pid_obj, drgn_object_program(res));
drgn_object_init(&pid_type_obj, drgn_object_program(res));
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
err = linux_helper_find_pid(&pid_obj, ns, pid);
if (err)
goto out;
libdrgn: remove struct drgn_object::prog We can get it via the type now. Signed-off-by: Omar Sandoval <osandov@osandov.com>
2020-08-27 19:26:39 +01:00
err = drgn_program_find_object(drgn_object_program(res), "PIDTYPE_PID",
NULL, DRGN_FIND_OBJECT_CONSTANT,
Translate find_task() helper (and dependencies) to C We'd like to be able to look up tasks by PID from libdrgn, but those helpers are written in Python. Translate them to C and add some thin bindings so we can use the same implementation from Python.
2019-10-25 23:23:02 +01:00
&pid_type_obj);
if (err)
goto out;
err = drgn_object_read_integer(&pid_type_obj, &pid_type);
if (err)
goto out;
err = linux_helper_pid_task(res, &pid_obj, pid_type.uvalue);
out:
drgn_object_deinit(&pid_type_obj);
drgn_object_deinit(&pid_obj);
return err;
}
libdrgn: add Linux kernel task iterator The thread API needs a way to iterate over all task_structs in the kernel. Previously, we translated the existing for_each_task helper, which supports iterating through specific PID namespaces by walking through the PID radix tree or PID hashtable. However, we don't need specific namespaces for the thread API, so we can instead use the much simpler linked lists of thread groups and threads. Signed-off-by: Kevin Svetlitski <svetlitski@fb.com>
2021-12-18 00:56:00 +00:00
struct drgn_error *
linux_helper_task_iterator_init(struct linux_helper_task_iterator *it,
struct drgn_program *prog)
{
struct drgn_error *err;
it->done = false;
drgn_object_init(&it->task, prog);
err = drgn_program_find_object(prog, "init_task", NULL,
DRGN_FIND_OBJECT_VARIABLE, &it->task);
if (err)
goto err;
it->task_struct_type = drgn_object_qualified_type(&it->task);
err = drgn_object_address_of(&it->task, &it->task);
if (err)
goto err;
err = drgn_object_read_unsigned(&it->task, &it->init_task_address);
if (err)
goto err;
it->thread_group_address = it->init_task_address;
return NULL;
err:
drgn_object_deinit(&it->task);
return err;
}
void linux_helper_task_iterator_deinit(struct linux_helper_task_iterator *it)
{
drgn_object_deinit(&it->task);
}
struct drgn_error *
linux_helper_task_iterator_next(struct linux_helper_task_iterator *it,
Implement a new API for representing threads Previously, drgn had no way to represent a thread – retrieving a stack trace (the only extant thread-specific operation) was achieved by requiring the user to directly provide a tid. This commit introduces the scaffolding for the design outlined in issue #92, and implements the corresponding methods for userspace core dumps, the live Linux kernel, and Linux kernel core dumps. Future work will build on top of this commit to support live userspace processes. Signed-off-by: Kevin Svetlitski <svetlitski@fb.com>
2021-11-19 00:46:59 +00:00
const struct drgn_object **ret)
libdrgn: add Linux kernel task iterator The thread API needs a way to iterate over all task_structs in the kernel. Previously, we translated the existing for_each_task helper, which supports iterating through specific PID namespaces by walking through the PID radix tree or PID hashtable. However, we don't need specific namespaces for the thread API, so we can instead use the much simpler linked lists of thread groups and threads. Signed-off-by: Kevin Svetlitski <svetlitski@fb.com>
2021-12-18 00:56:00 +00:00
{
if (it->done) {
*ret = NULL;
return NULL;
}
struct drgn_error *err;
struct drgn_object *task = &it->task;
err = drgn_object_member_dereference(task, task, "thread_group");
if (err)
return err;
err = drgn_object_member(task, task, "next");
if (err)
return err;
err = drgn_object_container_of(task, task, it->task_struct_type,
"thread_group");
if (err)
return err;
uint64_t task_address;
err = drgn_object_read_unsigned(task, &task_address);
if (err)
return err;
if (task_address == it->thread_group_address) {
err = drgn_object_member_dereference(task, task, "tasks");
if (err)
return err;
err = drgn_object_member(task, task, "next");
if (err)
return err;
err = drgn_object_container_of(task, task,
it->task_struct_type, "tasks");
if (err)
return err;
err = drgn_object_read_unsigned(task,
&it->thread_group_address);
if (err)
return err;
if (it->thread_group_address == it->init_task_address)
it->done = true;
}
*ret = task;
return NULL;
}
Reference in New Issue Copy Permalink