Introduces the function is_breg_access_indirect to determine whether a
memory access involving a DW_OP_breg* operation refers to the variable's
value directly or requires dereferencing the variable's type as a
pointer based on the DWARF expression. Previously, all breg based
accesses were assumed to directly access the variable's value
(is_pointer = false).
The is_breg_access_indirect function handles three cases:
1. Base register + offset only: (e.g., DW_OP_breg7 RSP+88) The
calculated address is the location of the variable. The access is
direct, so no type dereference is needed. Returns false.
2. Base register + offset, followed by other operations ending in
DW_OP_stack_value, including DW_OP_deref: (e.g., DW_OP_breg*,
DW_OP_deref, DW_OP_stack_value) The DWARF expression computes the
variable's value, but that value requires a dereference. The memory
access is fetching that value, so no type dereference is needed.
Returns false.
3. Base register + offset, followed only by DW_OP_stack_value: (e.g.,
DW_OP_breg13 R13+256, DW_OP_stack_value) This indicates the value at
the base + offset is the variable's value. Since this value is being
used as an address in the memory access, the variable's type is
treated as a pointer and requires a type dereference. Returns true.
The is_pointer argument passed to match_var_offset is now set by
is_breg_access_indirect for breg accesses.
There are more complex expressions that includes multiple operations and
may require additional handling, such as DW_OP_deref without a
DW_OP_stack_value, or including multiple base registers. They are less
common in the Linux kernel dwarf and are skipped in check_allowed_ops.
Signed-off-by: Zecheng Li <zecheng@google.com>
---
tools/perf/util/dwarf-aux.c | 38 ++++++++++++++++++++++++++++++++-----
1 file changed, 33 insertions(+), 5 deletions(-)
diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
index 920054425578..449bc9ad7aff 100644
--- a/tools/perf/util/dwarf-aux.c
+++ b/tools/perf/util/dwarf-aux.c
@@ -1423,6 +1423,34 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
return true;
}
+/**
+ * is_breg_access_indirect - Check if breg based access implies type
+ * dereference
+ * @ops: DWARF operations array
+ * @nops: Number of operations in @ops
+ *
+ * Returns true if the DWARF expression evaluates to the variable's
+ * value, so the memory access on that register needs type dereference.
+ * Returns false if the expression evaluates to the variable's address.
+ * This is called after check_allowed_ops.
+ */
+static bool is_breg_access_indirect(Dwarf_Op *ops, size_t nops)
+{
+ /* only the base register */
+ if (nops == 1)
+ return false;
+
+ if (nops == 2 && ops[1].atom == DW_OP_stack_value)
+ return true;
+
+ if (nops == 3 && (ops[1].atom == DW_OP_deref ||
+ ops[1].atom == DW_OP_deref_size) &&
+ ops[2].atom == DW_OP_stack_value)
+ return false;
+ /* unreachable, OP not supported */
+ return false;
+}
+
/* Only checks direct child DIEs in the given scope. */
static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
{
@@ -1451,7 +1479,7 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
if (data->is_fbreg && ops->atom == DW_OP_fbreg &&
check_allowed_ops(ops, nops) &&
match_var_offset(die_mem, data, data->offset, ops->number,
- /*is_pointer=*/false))
+ is_breg_access_indirect(ops, nops)))
return DIE_FIND_CB_END;
/* Only match with a simple case */
@@ -1463,11 +1491,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
/*is_pointer=*/true))
return DIE_FIND_CB_END;
- /* Local variables accessed by a register + offset */
+ /* variables accessed by a register + offset */
if (ops->atom == (DW_OP_breg0 + data->reg) &&
check_allowed_ops(ops, nops) &&
match_var_offset(die_mem, data, data->offset, ops->number,
- /*is_pointer=*/false))
+ is_breg_access_indirect(ops, nops)))
return DIE_FIND_CB_END;
} else {
/* pointer variables saved in a register 32 or above */
@@ -1477,11 +1505,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
/*is_pointer=*/true))
return DIE_FIND_CB_END;
- /* Local variables accessed by a register + offset */
+ /* variables accessed by a register + offset */
if (ops->atom == DW_OP_bregx && data->reg == ops->number &&
check_allowed_ops(ops, nops) &&
match_var_offset(die_mem, data, data->offset, ops->number2,
- /*is_poitner=*/false))
+ is_breg_access_indirect(ops, nops)))
return DIE_FIND_CB_END;
}
}
--
2.51.0.261.g7ce5a0a67e-googOn Mon, Aug 25, 2025 at 07:54:04PM +0000, Zecheng Li wrote:
> Introduces the function is_breg_access_indirect to determine whether a
> memory access involving a DW_OP_breg* operation refers to the variable's
> value directly or requires dereferencing the variable's type as a
> pointer based on the DWARF expression. Previously, all breg based
> accesses were assumed to directly access the variable's value
> (is_pointer = false).
>
> The is_breg_access_indirect function handles three cases:
>
> 1. Base register + offset only: (e.g., DW_OP_breg7 RSP+88) The
> calculated address is the location of the variable. The access is
> direct, so no type dereference is needed. Returns false.
I'm afraid there may be cases that the base register doesn't point to
the stack. In that case it may return true, right?
I think struct find_var_data already has 'is_fbreg' field. Maybe you
can add 'is_stack' or 'is_stack_reg' field if the target. Currently we
hardcoded X86_REG_SP but it should be arch-dependent.
>
> 2. Base register + offset, followed by other operations ending in
> DW_OP_stack_value, including DW_OP_deref: (e.g., DW_OP_breg*,
> DW_OP_deref, DW_OP_stack_value) The DWARF expression computes the
> variable's value, but that value requires a dereference. The memory
> access is fetching that value, so no type dereference is needed.
> Returns false.
>
> 3. Base register + offset, followed only by DW_OP_stack_value: (e.g.,
> DW_OP_breg13 R13+256, DW_OP_stack_value) This indicates the value at
> the base + offset is the variable's value. Since this value is being
> used as an address in the memory access, the variable's type is
> treated as a pointer and requires a type dereference. Returns true.
>
> The is_pointer argument passed to match_var_offset is now set by
> is_breg_access_indirect for breg accesses.
>
> There are more complex expressions that includes multiple operations and
> may require additional handling, such as DW_OP_deref without a
> DW_OP_stack_value, or including multiple base registers. They are less
> common in the Linux kernel dwarf and are skipped in check_allowed_ops.
>
> Signed-off-by: Zecheng Li <zecheng@google.com>
> ---
> tools/perf/util/dwarf-aux.c | 38 ++++++++++++++++++++++++++++++++-----
> 1 file changed, 33 insertions(+), 5 deletions(-)
>
> diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
> index 920054425578..449bc9ad7aff 100644
> --- a/tools/perf/util/dwarf-aux.c
> +++ b/tools/perf/util/dwarf-aux.c
> @@ -1423,6 +1423,34 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
> return true;
> }
>
> +/**
> + * is_breg_access_indirect - Check if breg based access implies type
> + * dereference
> + * @ops: DWARF operations array
> + * @nops: Number of operations in @ops
> + *
> + * Returns true if the DWARF expression evaluates to the variable's
> + * value, so the memory access on that register needs type dereference.
> + * Returns false if the expression evaluates to the variable's address.
> + * This is called after check_allowed_ops.
> + */
> +static bool is_breg_access_indirect(Dwarf_Op *ops, size_t nops)
> +{
> + /* only the base register */
> + if (nops == 1)
> + return false;
Then it could be like below:
if (nops == 1) {
int reg = reg_from_dwarf_op(ops);
return !(reg == DWARF_REG_FB || data->is_fbreg || reg == data->is_stack);
}
Thanks,
Namhyung
> +
> + if (nops == 2 && ops[1].atom == DW_OP_stack_value)
> + return true;
> +
> + if (nops == 3 && (ops[1].atom == DW_OP_deref ||
> + ops[1].atom == DW_OP_deref_size) &&
> + ops[2].atom == DW_OP_stack_value)
> + return false;
> + /* unreachable, OP not supported */
> + return false;
> +}
> +
> /* Only checks direct child DIEs in the given scope. */
> static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> {
> @@ -1451,7 +1479,7 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> if (data->is_fbreg && ops->atom == DW_OP_fbreg &&
> check_allowed_ops(ops, nops) &&
> match_var_offset(die_mem, data, data->offset, ops->number,
> - /*is_pointer=*/false))
> + is_breg_access_indirect(ops, nops)))
> return DIE_FIND_CB_END;
>
> /* Only match with a simple case */
> @@ -1463,11 +1491,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> /*is_pointer=*/true))
> return DIE_FIND_CB_END;
>
> - /* Local variables accessed by a register + offset */
> + /* variables accessed by a register + offset */
> if (ops->atom == (DW_OP_breg0 + data->reg) &&
> check_allowed_ops(ops, nops) &&
> match_var_offset(die_mem, data, data->offset, ops->number,
> - /*is_pointer=*/false))
> + is_breg_access_indirect(ops, nops)))
> return DIE_FIND_CB_END;
> } else {
> /* pointer variables saved in a register 32 or above */
> @@ -1477,11 +1505,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> /*is_pointer=*/true))
> return DIE_FIND_CB_END;
>
> - /* Local variables accessed by a register + offset */
> + /* variables accessed by a register + offset */
> if (ops->atom == DW_OP_bregx && data->reg == ops->number &&
> check_allowed_ops(ops, nops) &&
> match_var_offset(die_mem, data, data->offset, ops->number2,
> - /*is_poitner=*/false))
> + is_breg_access_indirect(ops, nops)))
> return DIE_FIND_CB_END;
> }
> }
> --
> 2.51.0.261.g7ce5a0a67e-goog
>
On Thu, Aug 28, 2025 at 3:19 AM Namhyung Kim <namhyung@kernel.org> wrote:
>
> On Mon, Aug 25, 2025 at 07:54:04PM +0000, Zecheng Li wrote:
> > Introduces the function is_breg_access_indirect to determine whether a
> > memory access involving a DW_OP_breg* operation refers to the variable's
> > value directly or requires dereferencing the variable's type as a
> > pointer based on the DWARF expression. Previously, all breg based
> > accesses were assumed to directly access the variable's value
> > (is_pointer = false).
> >
> > The is_breg_access_indirect function handles three cases:
> >
> > 1. Base register + offset only: (e.g., DW_OP_breg7 RSP+88) The
> > calculated address is the location of the variable. The access is
> > direct, so no type dereference is needed. Returns false.
>
> I'm afraid there may be cases that the base register doesn't point to
> the stack. In that case it may return true, right?
Hi Namhyung,
In this case, the DWARF specification for a DW_OP_breg* operation is
to always calculate a memory address. So, even if the base register
isn't the stack pointer, the expression still resolves to the
variable's location, meaning the access is direct (is_pointer =
false).
> I think struct find_var_data already has 'is_fbreg' field. Maybe you
> can add 'is_stack' or 'is_stack_reg' field if the target. Currently we
> hardcoded X86_REG_SP but it should be arch-dependent.
Therefore we don't need to check if the register is a stack or frame base.
> >
> > 2. Base register + offset, followed by other operations ending in
> > DW_OP_stack_value, including DW_OP_deref: (e.g., DW_OP_breg*,
> > DW_OP_deref, DW_OP_stack_value) The DWARF expression computes the
> > variable's value, but that value requires a dereference. The memory
> > access is fetching that value, so no type dereference is needed.
> > Returns false.
> >
> > 3. Base register + offset, followed only by DW_OP_stack_value: (e.g.,
> > DW_OP_breg13 R13+256, DW_OP_stack_value) This indicates the value at
> > the base + offset is the variable's value. Since this value is being
> > used as an address in the memory access, the variable's type is
> > treated as a pointer and requires a type dereference. Returns true.
> >
> > The is_pointer argument passed to match_var_offset is now set by
> > is_breg_access_indirect for breg accesses.
> >
> > There are more complex expressions that includes multiple operations and
> > may require additional handling, such as DW_OP_deref without a
> > DW_OP_stack_value, or including multiple base registers. They are less
> > common in the Linux kernel dwarf and are skipped in check_allowed_ops.
> >
> > Signed-off-by: Zecheng Li <zecheng@google.com>
> > ---
> > tools/perf/util/dwarf-aux.c | 38 ++++++++++++++++++++++++++++++++-----
> > 1 file changed, 33 insertions(+), 5 deletions(-)
> >
> > diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
> > index 920054425578..449bc9ad7aff 100644
> > --- a/tools/perf/util/dwarf-aux.c
> > +++ b/tools/perf/util/dwarf-aux.c
> > @@ -1423,6 +1423,34 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
> > return true;
> > }
> >
> > +/**
> > + * is_breg_access_indirect - Check if breg based access implies type
> > + * dereference
> > + * @ops: DWARF operations array
> > + * @nops: Number of operations in @ops
> > + *
> > + * Returns true if the DWARF expression evaluates to the variable's
> > + * value, so the memory access on that register needs type dereference.
> > + * Returns false if the expression evaluates to the variable's address.
> > + * This is called after check_allowed_ops.
> > + */
> > +static bool is_breg_access_indirect(Dwarf_Op *ops, size_t nops)
> > +{
> > + /* only the base register */
> > + if (nops == 1)
> > + return false;
>
> Then it could be like below:
>
> if (nops == 1) {
> int reg = reg_from_dwarf_op(ops);
> return !(reg == DWARF_REG_FB || data->is_fbreg || reg == data->is_stack);
> }
>
> Thanks,
> Namhyung
>
> > +
> > + if (nops == 2 && ops[1].atom == DW_OP_stack_value)
> > + return true;
> > +
> > + if (nops == 3 && (ops[1].atom == DW_OP_deref ||
> > + ops[1].atom == DW_OP_deref_size) &&
> > + ops[2].atom == DW_OP_stack_value)
> > + return false;
> > + /* unreachable, OP not supported */
> > + return false;
> > +}
> > +
> > /* Only checks direct child DIEs in the given scope. */
> > static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > {
> > @@ -1451,7 +1479,7 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > if (data->is_fbreg && ops->atom == DW_OP_fbreg &&
> > check_allowed_ops(ops, nops) &&
> > match_var_offset(die_mem, data, data->offset, ops->number,
> > - /*is_pointer=*/false))
> > + is_breg_access_indirect(ops, nops)))
> > return DIE_FIND_CB_END;
> >
> > /* Only match with a simple case */
> > @@ -1463,11 +1491,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > /*is_pointer=*/true))
> > return DIE_FIND_CB_END;
> >
> > - /* Local variables accessed by a register + offset */
> > + /* variables accessed by a register + offset */
> > if (ops->atom == (DW_OP_breg0 + data->reg) &&
> > check_allowed_ops(ops, nops) &&
> > match_var_offset(die_mem, data, data->offset, ops->number,
> > - /*is_pointer=*/false))
> > + is_breg_access_indirect(ops, nops)))
> > return DIE_FIND_CB_END;
> > } else {
> > /* pointer variables saved in a register 32 or above */
> > @@ -1477,11 +1505,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > /*is_pointer=*/true))
> > return DIE_FIND_CB_END;
> >
> > - /* Local variables accessed by a register + offset */
> > + /* variables accessed by a register + offset */
> > if (ops->atom == DW_OP_bregx && data->reg == ops->number &&
> > check_allowed_ops(ops, nops) &&
> > match_var_offset(die_mem, data, data->offset, ops->number2,
> > - /*is_poitner=*/false))
> > + is_breg_access_indirect(ops, nops)))
> > return DIE_FIND_CB_END;
> > }
> > }
> > --
> > 2.51.0.261.g7ce5a0a67e-goog
> >
Hello,
On Thu, Aug 28, 2025 at 02:36:32PM -0400, Zecheng Li wrote:
> On Thu, Aug 28, 2025 at 3:19 AM Namhyung Kim <namhyung@kernel.org> wrote:
> >
> > On Mon, Aug 25, 2025 at 07:54:04PM +0000, Zecheng Li wrote:
> > > Introduces the function is_breg_access_indirect to determine whether a
> > > memory access involving a DW_OP_breg* operation refers to the variable's
> > > value directly or requires dereferencing the variable's type as a
> > > pointer based on the DWARF expression. Previously, all breg based
> > > accesses were assumed to directly access the variable's value
> > > (is_pointer = false).
> > >
> > > The is_breg_access_indirect function handles three cases:
> > >
> > > 1. Base register + offset only: (e.g., DW_OP_breg7 RSP+88) The
> > > calculated address is the location of the variable. The access is
> > > direct, so no type dereference is needed. Returns false.
> >
> > I'm afraid there may be cases that the base register doesn't point to
> > the stack. In that case it may return true, right?
>
> Hi Namhyung,
>
> In this case, the DWARF specification for a DW_OP_breg* operation is
> to always calculate a memory address. So, even if the base register
> isn't the stack pointer, the expression still resolves to the
> variable's location, meaning the access is direct (is_pointer =
> false).
I've re-read the DWARF spec and I think you're right. :)
>
> > I think struct find_var_data already has 'is_fbreg' field. Maybe you
> > can add 'is_stack' or 'is_stack_reg' field if the target. Currently we
> > hardcoded X86_REG_SP but it should be arch-dependent.
>
> Therefore we don't need to check if the register is a stack or frame base.
Fair enough.
>
> > >
> > > 2. Base register + offset, followed by other operations ending in
> > > DW_OP_stack_value, including DW_OP_deref: (e.g., DW_OP_breg*,
> > > DW_OP_deref, DW_OP_stack_value) The DWARF expression computes the
> > > variable's value, but that value requires a dereference. The memory
> > > access is fetching that value, so no type dereference is needed.
> > > Returns false.
> > >
> > > 3. Base register + offset, followed only by DW_OP_stack_value: (e.g.,
> > > DW_OP_breg13 R13+256, DW_OP_stack_value) This indicates the value at
> > > the base + offset is the variable's value. Since this value is being
> > > used as an address in the memory access, the variable's type is
> > > treated as a pointer and requires a type dereference. Returns true.
> > >
> > > The is_pointer argument passed to match_var_offset is now set by
> > > is_breg_access_indirect for breg accesses.
> > >
> > > There are more complex expressions that includes multiple operations and
> > > may require additional handling, such as DW_OP_deref without a
> > > DW_OP_stack_value, or including multiple base registers. They are less
> > > common in the Linux kernel dwarf and are skipped in check_allowed_ops.
> > >
> > > Signed-off-by: Zecheng Li <zecheng@google.com>
This also improved the data quality!
Reviewed-by: Namhyung Kim <namhyung@kernel.org>
Thanks,
Namhyung
> > > ---
> > > tools/perf/util/dwarf-aux.c | 38 ++++++++++++++++++++++++++++++++-----
> > > 1 file changed, 33 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
> > > index 920054425578..449bc9ad7aff 100644
> > > --- a/tools/perf/util/dwarf-aux.c
> > > +++ b/tools/perf/util/dwarf-aux.c
> > > @@ -1423,6 +1423,34 @@ static bool match_var_offset(Dwarf_Die *die_mem, struct find_var_data *data,
> > > return true;
> > > }
> > >
> > > +/**
> > > + * is_breg_access_indirect - Check if breg based access implies type
> > > + * dereference
> > > + * @ops: DWARF operations array
> > > + * @nops: Number of operations in @ops
> > > + *
> > > + * Returns true if the DWARF expression evaluates to the variable's
> > > + * value, so the memory access on that register needs type dereference.
> > > + * Returns false if the expression evaluates to the variable's address.
> > > + * This is called after check_allowed_ops.
> > > + */
> > > +static bool is_breg_access_indirect(Dwarf_Op *ops, size_t nops)
> > > +{
> > > + /* only the base register */
> > > + if (nops == 1)
> > > + return false;
> >
> > Then it could be like below:
> >
> > if (nops == 1) {
> > int reg = reg_from_dwarf_op(ops);
> > return !(reg == DWARF_REG_FB || data->is_fbreg || reg == data->is_stack);
> > }
> >
> > Thanks,
> > Namhyung
> >
> > > +
> > > + if (nops == 2 && ops[1].atom == DW_OP_stack_value)
> > > + return true;
> > > +
> > > + if (nops == 3 && (ops[1].atom == DW_OP_deref ||
> > > + ops[1].atom == DW_OP_deref_size) &&
> > > + ops[2].atom == DW_OP_stack_value)
> > > + return false;
> > > + /* unreachable, OP not supported */
> > > + return false;
> > > +}
> > > +
> > > /* Only checks direct child DIEs in the given scope. */
> > > static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > > {
> > > @@ -1451,7 +1479,7 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > > if (data->is_fbreg && ops->atom == DW_OP_fbreg &&
> > > check_allowed_ops(ops, nops) &&
> > > match_var_offset(die_mem, data, data->offset, ops->number,
> > > - /*is_pointer=*/false))
> > > + is_breg_access_indirect(ops, nops)))
> > > return DIE_FIND_CB_END;
> > >
> > > /* Only match with a simple case */
> > > @@ -1463,11 +1491,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > > /*is_pointer=*/true))
> > > return DIE_FIND_CB_END;
> > >
> > > - /* Local variables accessed by a register + offset */
> > > + /* variables accessed by a register + offset */
> > > if (ops->atom == (DW_OP_breg0 + data->reg) &&
> > > check_allowed_ops(ops, nops) &&
> > > match_var_offset(die_mem, data, data->offset, ops->number,
> > > - /*is_pointer=*/false))
> > > + is_breg_access_indirect(ops, nops)))
> > > return DIE_FIND_CB_END;
> > > } else {
> > > /* pointer variables saved in a register 32 or above */
> > > @@ -1477,11 +1505,11 @@ static int __die_find_var_reg_cb(Dwarf_Die *die_mem, void *arg)
> > > /*is_pointer=*/true))
> > > return DIE_FIND_CB_END;
> > >
> > > - /* Local variables accessed by a register + offset */
> > > + /* variables accessed by a register + offset */
> > > if (ops->atom == DW_OP_bregx && data->reg == ops->number &&
> > > check_allowed_ops(ops, nops) &&
> > > match_var_offset(die_mem, data, data->offset, ops->number2,
> > > - /*is_poitner=*/false))
> > > + is_breg_access_indirect(ops, nops)))
> > > return DIE_FIND_CB_END;
> > > }
> > > }
> > > --
> > > 2.51.0.261.g7ce5a0a67e-goog
> > >
© 2016 - 2025 Red Hat, Inc.