drgn/libdrgn/object.h
Omar Sandoval 243f6fb7d5 libdrgn: support value objects with >64-bit integer types
The Linux kernel's struct task_struct on AArch64 contains an array of
__uint128_t:

  >>> task = find_task(prog, 1)
  >>> task.type_
  struct task_struct *
  >>> task.thread.type_
  struct thread_struct {
          struct cpu_context cpu_context;
          struct {
                  unsigned long tp_value;
                  unsigned long tp2_value;
                  struct user_fpsimd_state fpsimd_state;
          } uw;
          enum fp_type fp_type;
          unsigned int fpsimd_cpu;
          void *sve_state;
          void *sme_state;
          unsigned int vl[2];
          unsigned int vl_onexec[2];
          unsigned long fault_address;
          unsigned long fault_code;
          struct debug_info debug;
          struct ptrauth_keys_user keys_user;
          struct ptrauth_keys_kernel keys_kernel;
          u64 mte_ctrl;
          u64 sctlr_user;
          u64 svcr;
          u64 tpidr2_el0;
  }
  >>> task.thread.uw.fpsimd_state.type_
  struct user_fpsimd_state {
          __int128 unsigned vregs[32];
          __u32 fpsr;
          __u32 fpcr;
          __u32 __reserved[2];
  }

As a result, printing a task_struct fails:

  >>> task
  Traceback (most recent call last):
    File "<console>", line 1, in <module>
    File "/host/home/osandov/repos/drgn3/drgn/cli.py", line 140, in _displayhook
      text = value.format_(columns=shutil.get_terminal_size((0, 0)).columns)
  NotImplementedError: integer values larger than 64 bits are not yet supported

PR #311 suggested treating >64-bit integers as byte arrays for now; I
tried an alternate hack of handling >64-bit integers only in the
pretty-printing code. Both of these had issues, though.

Instead, let's push >64-bit integer support a little further and allow
storing "big integer" value objects. We still don't support any
operations on them, so this still doesn't complete #170. We store the
raw bytes of the value for now, but we'll probably change this if we add
support for operations (e.g., to store the value as an mp_limb_t array
for GMP). We also print >64-bit integer types in hexadecimal for
simplicity. This is inconsistent with the existing behavior of printing
in decimal, but more readable. In the future, we might want to add
heuristics to decide when to print in decimal vs hexadecimal for all
sizes.

Closes #311.

Signed-off-by: Omar Sandoval <osandov@osandov.com>
2023-08-02 14:21:46 -07:00

402 lines
12 KiB
C

// Copyright (c) Meta Platforms, Inc. and affiliates.
// SPDX-License-Identifier: LGPL-2.1-or-later
/**
* @file
*
* Object internals.
*
* See @ref ObjectInternals.
*/
#ifndef DRGN_OBJECT_H
#define DRGN_OBJECT_H
#include <stdlib.h>
#include <string.h>
#include "drgn.h"
#include "type.h"
/**
* @ingroup Internals
*
* @defgroup ObjectInternals Objects
*
* Object internals.
*
* This provides the language-agnostic part of operator implementations. The
* operators have defined behavior for various cases where C is undefined or
* implementation-defined (e.g., signed arithmetic is modular, signed bitwise
* operators operate on the two's complement representation, right shifts are
* arithmetic).
*
* @{
*/
/** Allocate a zero-initialized @ref drgn_value. */
static inline bool drgn_value_zalloc(uint64_t size, union drgn_value *value_ret,
char **buf_ret)
{
if (size <= sizeof(value_ret->ibuf)) {
memset(value_ret->ibuf, 0, sizeof(value_ret->ibuf));
*buf_ret = value_ret->ibuf;
} else {
if (size > SIZE_MAX)
return false;
char *buf = calloc(1, size);
if (!buf)
return false;
value_ret->bufp = *buf_ret = buf;
}
return true;
}
/**
* Get whether an object is zero.
*
* For scalars, this is true iff its value is zero. For structures, unions, and
* classes, this is true iff all of its members are zero. For arrays, this is
* true iff all of its elements are zero. Note that this ignores padding.
*/
struct drgn_error *drgn_object_is_zero(const struct drgn_object *obj,
bool *ret);
/** Type-related fields from @ref drgn_object. */
struct drgn_object_type {
struct drgn_type *type;
/* Cached underlying type of @c type. */
struct drgn_type *underlying_type;
uint64_t bit_size;
enum drgn_qualifiers qualifiers;
enum drgn_object_encoding encoding;
bool is_bit_field;
bool little_endian;
};
/** Convert a @ref drgn_object_type to a @ref drgn_qualified_type. */
static inline struct drgn_qualified_type
drgn_object_type_qualified(const struct drgn_object_type *type)
{
return (struct drgn_qualified_type){
.type = type->type,
.qualifiers = type->qualifiers,
};
}
/**
* Type of an operand or operator result.
*
* This is basically @ref drgn_qualified_type plus a bit field size and cached
* underlying type.
*/
struct drgn_operand_type {
struct drgn_type *type;
enum drgn_qualifiers qualifiers;
struct drgn_type *underlying_type;
uint64_t bit_field_size;
};
/** Get the @ref drgn_operand_type of a @ref drgn_object. */
static inline struct drgn_operand_type
drgn_object_operand_type(const struct drgn_object *obj)
{
return (struct drgn_operand_type){
.type = obj->type,
.qualifiers = obj->qualifiers,
.underlying_type = drgn_underlying_type(obj->type),
.bit_field_size = obj->is_bit_field ? obj->bit_size : 0,
};
}
/**
* Deinitialize the value of a @ref drgn_object and reinitialize the kind and
* type fields.
*/
static inline void drgn_object_reinit(struct drgn_object *obj,
const struct drgn_object_type *type,
enum drgn_object_kind kind)
{
drgn_object_deinit(obj);
obj->type = type->type;
obj->qualifiers = type->qualifiers;
obj->bit_size = type->bit_size;
obj->encoding = type->encoding;
obj->is_bit_field = type->is_bit_field;
obj->little_endian = type->little_endian;
obj->kind = kind;
}
/**
* Compute the type-related fields of a @ref drgn_object from a @ref
* drgn_qualified_type and a bit field size.
*/
struct drgn_error *
drgn_object_type(struct drgn_qualified_type qualified_type,
uint64_t bit_field_size, struct drgn_object_type *ret);
/**
* Like @ref drgn_object_set_signed() but @ref drgn_object_type() was already
* called and the type is already known to be a signed integer type.
*/
struct drgn_error *
drgn_object_set_signed_internal(struct drgn_object *res,
const struct drgn_object_type *type,
int64_t svalue);
/**
* Like @ref drgn_object_set_unsigned() but @ref drgn_object_type() was already
* called and the type is already known to be an unsigned integer type.
*/
struct drgn_error *
drgn_object_set_unsigned_internal(struct drgn_object *res,
const struct drgn_object_type *type,
uint64_t uvalue);
/**
* Like @ref drgn_object_set_float() but @ref drgn_object_type() was already
* called and the type is already known to be a floating-point type.
*/
struct drgn_error *
drgn_object_set_float_internal(struct drgn_object *res,
const struct drgn_object_type *type,
double fvalue);
/**
* Like @ref drgn_object_set_from_buffer() but @ref drgn_object_type() was
* already called and the bounds of the buffer have already been checked.
*/
struct drgn_error *
drgn_object_set_from_buffer_internal(struct drgn_object *res,
const struct drgn_object_type *type,
const void *buf, uint64_t bit_offset);
/**
* Like @ref drgn_object_set_reference() but @ref drgn_object_type() was already
* called.
*/
struct drgn_error *
drgn_object_set_reference_internal(struct drgn_object *res,
const struct drgn_object_type *type,
uint64_t address, uint64_t bit_offset);
/**
* Binary operator implementation.
*
* Operator implementations with this type convert @p lhs and @p rhs to @p
* op_type, apply the operator, and store the result in @p res.
*
* @param[out] res Operator result. May be the same as @p lhs and/or @p rhs.
* @param[in] op_type Result type.
* @param[in] lhs Operator left hand side.
* @param[in] rhs Operator right hand side.
* @return @c NULL on success, non-@c NULL on error. @p res is not modified on
* error.
*/
typedef struct drgn_error *
drgn_binary_op_impl(struct drgn_object *res,
const struct drgn_operand_type *op_type,
const struct drgn_object *lhs,
const struct drgn_object *rhs);
/**
* Shift operator implementation.
*
* Operator implementations with this type convert @p lhs to @p lhs_type and @p
* rhs to @p rhs_type and store the result in @p res.
*
* @param[out] res Operator result. May be the same as @p lhs and/or @p rhs.
* @param[in] lhs Operator left hand side.
* @param[in] lhs_type Type of left hand side and result.
* @param[in] rhs Operator right hand side.
* @param[in] rhs_type Type of right hand side.
* @return @c NULL on success, non-@c NULL on error. @p res is not modified on
* error.
*/
typedef struct drgn_error *
drgn_shift_op_impl(struct drgn_object *res,
const struct drgn_object *lhs,
const struct drgn_operand_type *lhs_type,
const struct drgn_object *rhs,
const struct drgn_operand_type *rhs_type);
/**
* Unary operator implementation.
*
* Operator implementations with this type convert @p obj to @p op_type and
* store the result in @p res.
*
* @param[out] res Operator result. May be the same as @p obj.
* @param[in] op_type Result type.
* @param[in] obj Operand.
* @return @c NULL on success, non-@c NULL on error. @p res is not modified on
* error.
*/
typedef struct drgn_error *
drgn_unary_op_impl(struct drgn_object *res,
const struct drgn_operand_type *op_type,
const struct drgn_object *obj);
/**
* Implement addition for signed, unsigned, and floating-point objects.
*
* Integer results are reduced modulo 2^width.
*/
drgn_binary_op_impl drgn_op_add_impl;
/**
* Implement subtraction for signed, unsigned, and floating-point objects.
*
* Integer results are reduced modulo 2^width.
*/
drgn_binary_op_impl drgn_op_sub_impl;
/**
* Implement multiplication for signed, unsigned, and floating-point objects.
*
* Integer results are reduced modulo 2^width.
*/
drgn_binary_op_impl drgn_op_mul_impl;
/**
* Implement division for signed, unsigned, and floating-point objects.
*
* Integer results are truncated towards zero. A @ref DRGN_ERROR_ZERO_DIVISION
* error is returned if @p rhs is zero.
*/
drgn_binary_op_impl drgn_op_div_impl;
/**
* Implement modulo for signed and unsigned objects.
*
* The result has the sign of the dividend. A @ref DRGN_ERROR_ZERO_DIVISION
* error is returned if @p rhs is zero.
*/
drgn_binary_op_impl drgn_op_mod_impl;
/**
* Implement left shift for signed and unsigned objects.
*
* For signed integers, this acts on the two's complement representation. The
* result is reduced modulo 2^width. In particular, if @p rhs is greater than
* the width of the result, then the result is zero. An error is returned if @p
* rhs is negative.
*/
drgn_shift_op_impl drgn_op_lshift_impl;
/**
* Implement right shift for signed and unsigned objects.
*
* For signed integers, this is an arithmetic shift. For unsigned integers, it
* is logical. The result is reduced modulo 2^width. In particular, if @p rhs is
* greater than the width of the result, then the result is zero. An error is
* returned if @p rhs is negative.
*/
drgn_shift_op_impl drgn_op_rshift_impl;
/**
* Implement bitwise and for signed and unsigned objects.
*
* For signed integers, this acts on the two's complement representation.
*/
drgn_binary_op_impl drgn_op_and_impl;
/**
* Implement bitwise or for signed and unsigned objects.
*
* For signed integers, this acts on the two's complement representation.
*/
drgn_binary_op_impl drgn_op_or_impl;
/**
* Implement bitwise xor for signed and unsigned objects.
*
* For signed integers, this acts on the two's complement representation.
*/
drgn_binary_op_impl drgn_op_xor_impl;
/**
* Implement the unary plus operator for signed, unsigned, and floating-point
* objects.
*
* This converts @p obj without otherwise changing the value.
*/
drgn_unary_op_impl drgn_op_pos_impl;
/**
* Implement negation for signed, unsigned, and floating-point objects.
*
* Integer results are reduced modulo 2^width.
*/
drgn_unary_op_impl drgn_op_neg_impl;
/**
* Implement bitwise negation for signed and unsigned objects.
*
* For signed integers, this acts on the two's complement representation.
*/
drgn_unary_op_impl drgn_op_not_impl;
/**
* Implement object type casting.
*
* If @p obj_type is a pointer type and @c obj is a buffer, then the reference
* address of @p obj is used.
*/
struct drgn_error *drgn_op_cast(struct drgn_object *res,
struct drgn_qualified_type qualified_type,
const struct drgn_object *obj,
const struct drgn_operand_type *obj_type);
/**
* Implement object comparison for signed, unsigned, and floating-point objects.
*
* This converts @p lhs and @p rhs to @p type before comparing.
*/
struct drgn_error *drgn_op_cmp_impl(const struct drgn_object *lhs,
const struct drgn_object *rhs,
const struct drgn_operand_type *op_type,
int *ret);
/**
* Implement object comparison for pointers and reference buffer objects.
*
* When comparing reference buffer objects, their address is used.
*/
struct drgn_error *drgn_op_cmp_pointers(const struct drgn_object *lhs,
const struct drgn_object *rhs,
int *ret);
/**
* Implement pointer arithmetic.
*
* This converts @p ptr to @p op_type, adds or subtracts
* <tt>index * referenced_size</tt>, and stores the result in @p res.
*
* @param[out] res Operator result. May be the same as @p ptr or @p index.
* @param[in] op_type Result type.
* @param[in] referenced_size Size of the object pointed to by @p ptr.
* @param[in] negate Subtract @p index instead of adding.
* @param[in] ptr Pointer.
* @param[in] index Value to add to/subtract from pointer.
* @return @c NULL on success, non-@c NULL on error. @p res is not modified on
* error.
*/
struct drgn_error *
drgn_op_add_to_pointer(struct drgn_object *res,
const struct drgn_operand_type *op_type,
uint64_t referenced_size, bool negate,
const struct drgn_object *ptr,
const struct drgn_object *index);
/**
* Implement pointer subtraction.
*
* This stores <tt>(lhs - rhs) / referenced_size</tt> in @p res.
*
* @param[out] res Operator result. May be the same as @p lhs and/or @p rhs.
* @param[in] referenced_size Size of the object pointed to by @p lhs and @p
* rhs.
* @param[in] op_type Result type. Must be a signed integer type.
* @param[in] lhs Operator left hand side.
* @param[in] rhs Operator right hand side.
* @return @c NULL on success, non-@c NULL on error. @p res is not modified on
* error.
*/
struct drgn_error *drgn_op_sub_pointers(struct drgn_object *res,
const struct drgn_operand_type *op_type,
uint64_t referenced_size,
const struct drgn_object *lhs,
const struct drgn_object *rhs);
/** @} */
#endif /* DRGN_OBJECT_H */