This reverts commit 7c5e35b63b.
Dropped since it could not cope with recursive types. A new attempt will
be made on 1.24.
Signed-off-by: Douglas RAILLARD <douglas.raillard@arm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Allow making inner struct enums and union anonymous, so that when using
-E to expand types we don't end up with multiple definitions for
expanded inner structs, allowing the resulting expanded struct to be
compilable.
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
[ Applied it manually to cover some fuzz due to other patches ]
Link: https://lore.kernel.org/all/20211019100724.325570-2-douglas.raillard@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Now we know what that 0x4a thing is:
$ pahole ~/c/split/foo.o
die__process: DW_TAG_compile_unit, DW_TAG_type_unit or DW_TAG_partial_unit expected got skeleton_unit (0x4a)!
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Remove the global variable and turn it into a member in struct
conf_fprintf, so that it can be used by other parts of the code.
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
Cc: dwarves@vger.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We just need to record if we printed it for a member and if so, deduce
that from the number of spaces left to print before the end of line
comment (offset, size).
Fixes: a59459bb80 ("fprintf: Account inline type __aligned__ member types for spacing")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This code:
struct X {
struct {
} __attribute__((foo)) x __attribute__((bar));
}
Was wrongly printed as:
struct X {
struct {
} x __attribute__((foo)) __attribute__((bar));
}
This unfortunately matters a lot, since "bar" is suppose to apply to
"x", but "foo" to typeof(x). In the wrong form, both apply to "x",
leading to e.g. incorrect layout for __aligned__ attribute.
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Systems such as Alpine Linux don't have support for obtaining _SC_LEVEL1_DCACHE_LINESIZE
via sysconf(), provide altenative method reading it from sysfs.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Another simplification made possible by using a plain char string
instead of string_t, that was only needed in the core as prep work
for CTF encoding.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Another simplification made possible by using a plain char string
instead of string_t, that was only needed in the core as prep work
for CTF encoding.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Another simplification made possible by using a plain char string
instead of string_t, that was only needed in the core as prep work
for CTF encoding.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Another simplification made possible by using a plain char string
instead of string_t, that was only needed in the core as prep work
for CTF encoding.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
For the threaded code we want to access strings in tags at the same time
that the string table may grow in another thread making the previous
pointer invalid, so, to avoid excessive locking, use plain strings.
The way the tools work will either consume the just produced CU straight
away or keep just one copy of each data structure when we keep all CUs
in memory, so lets try stopping using strings_t for strings.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
With the conversion of ->name members to plain char strings, no need
to use 'cu' to get the old string_t index and find the per-cu string
table.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
With the conversion of ->name members to plain char strings, no need
to use 'cu' to get the old string_t index and find the per-cu string
table.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
For the threaded code we want to access strings in tags at the same time
that the string table may grow in another thread making the previous
pointer invalid, so, to avoid excessive locking, use plain strings.
The way the tools work will either consume the just produced CU straight
away or keep just one copy of each data structure when we keep all CUs
in memory, so lets try stopping using strings_t for strings.
For the namespace->name case we get the bonus of removing another
user of dwarves__active_loader.
This covers unions, enums, structs and classes.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
For the threaded code we want to access strings in tags at the same time
that the string table may grow in another thread making the previous
pointer invalid, so, to avoid excessive locking, use plain strings.
The way the tools work will either consume the just produced CU straight
away or keep just one copy of each data structure when we keep all CUs
in memory, so lets try stopping using strings_t for strings.
For the class_member->name case we get the bonus of removing another
user of dwarves__active_loader.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
For the threaded code we want to access strings in tags at the same time
that the string table may grow in another thread making the previous
pointer invalid, so, to avoid excessive locking, use plain strings.
The way the tools work will either consume the just produced CU straight
away or keep just one copy of each data structure when we keep all CUs
in memory, so lets try stopping using strings_t for strings.
For the parameter->name case we get the bonus of removing a user of
dwarves__active_loader.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
For the threaded code we want to access strings in tags at the same time
that the string table may grow in another thread making the previous
pointer invalid, so, to avoid excessive locking, use plain strings.
The way the tools work will either consume the just produced CU straight
away or keep just one copy of each data structure when we keep all CUs
in memory, so lets try stopping using strings_t for strings.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We don't really reconstruct source code for FORTRAN, we just print it as
if it was C:
$ pahole examples/fortran95/derived-type.debug
struct bar {
integer(kind=4) c; /* 0 4 */
real(kind=4) d; /* 4 4 */
/* size: 8, cachelines: 1, members: 2 */
/* last cacheline: 8 bytes */
};
struct foo {
real(kind=4) a; /* 0 4 */
struct bar x; /* 4 8 */
string b[7]; /* 12 7 */
/* size: 20, cachelines: 1, members: 3 */
/* padding: 1 */
/* last cacheline: 20 bytes */
};
$
This comes from GCC build tests:
$ readelf -wi examples/fortran95/derived-type.debug | grep Fortran -A2
<9c> DW_AT_producer : (indirect string, offset: 0x1fb): GNU Fortran2008 10.2.1 20200728 [revision c0438ced53bcf57e4ebb1c38c226e41571aca892] -mtune=generic -march=x86-64 -g -fno-stack-protector -J /home/vries/gdb_versions/devel/build/gdb/testsuite/outputs/gdb.fortran/derived-type -fintrinsic-modules-path /usr/lib64/gcc/x86_64-suse-linux/10/finclude -fpre-include=/usr/include/finclude/math-vector-fortran.h
<a0> DW_AT_language : 14 (Fortran 95)
<a1> DW_AT_identifier_case: 2 (down_case)
<a2> DW_AT_name : (indirect string, offset: 0x365): /home/vries/gdb_versions/devel/src/gdb/testsuite/gdb.fortran/derived-type.f90
[acme@five pahole]$ readelf -wi examples/fortran95/derived-type.debug | grep DW_TAG_string_type -A2
<1><122>: Abbrev Number: 6 (DW_TAG_string_type)
<123> DW_AT_byte_size : 7
$
Now lets see whats more that is there segfaulting pahole, but for now I
think I don't have any segfaults, so just wait a bit for Hao to submit
the patch to selectively encode the per-cpu variables in BTF and then
cut v1.18.
Reported-by: Tom de Vries
Bugtracker: https://github.com/acmel/dwarves/issues/9
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Addresses this warning:
[ 26%] Building C object CMakeFiles/dwarves.dir/dwarves_fprintf.c.o
/home/acme/git/pahole/dwarves_fprintf.c: In function ‘type__fprintf’:
/home/acme/git/pahole/dwarves_fprintf.c:709:47: error: ‘%s’ directive output may be truncated writing up to 257 bytes into a region of size 256 [-Werror=format-truncation=]
709 | snprintf(namebfptr, sizeof(namebfptr), "* %s", name);
| ^~
/home/acme/git/pahole/dwarves_fprintf.c:709:5: note: ‘snprintf’ output between 3 and 260 bytes into a destination of size 258
709 | snprintf(namebfptr, sizeof(namebfptr), "* %s", name);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
That was affecting --suppress_aligned_attribute, detected using the
'btfdiff' utility, that uses pahole with DWARF and BTF and compares
their outputs, and since we don't have this in BTF, we use
--suppress_aligned_attribute with '-F dwarf'.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
As there are cases where we jump to the type not found label without
having done the full copy of *conf to tconf, so at least the
type_spacing needs to have been set.
This addresses this coverity report entry:
Error: UNINIT (CWE-457): [#def23]
dwarves-1.13/dwarves_fprintf.c:600: var_decl: Declaring variable "tconf" without initializer.
dwarves-1.13/dwarves_fprintf.c:779: uninit_use_in_call: Using uninitialized value "tconf.type_spacing" when calling "fprintf".
# 777| return printed;
# 778| out_type_not_found:
# 779|-> printed = fprintf(fp, "%-*s %s", tconf.type_spacing, "<ERROR>", name);
# 780| goto out;
# 781| }
Reported-by: William Cohen <wcohen@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In some exit paths we were accessing tconf before we had copied it from
conf, and we also were losing track of the original type that could have
been expanded and where we bumped the recursivity level member, so just
store that original type and if it is set, decrement its recursivity
level.
This addresses these coverity report entries:
Error: UNINIT (CWE-457): [#def21]
dwarves-1.13/dwarves_fprintf.c:600: var_decl: Declaring variable "tconf" without initializer.
dwarves-1.13/dwarves_fprintf.c:774: uninit_use: Using uninitialized value "tconf.expand_types".
# 772| }
# 773| out:
# 774|-> if (tconf.expand_types)
# 775| --type->recursivity_level;
# 776|
Error: FORWARD_NULL (CWE-476): [#def22]
dwarves-1.13/dwarves_fprintf.c:605: var_compare_op: Comparing "type" to null implies that "type" might be null.
dwarves-1.13/dwarves_fprintf.c:775: var_deref_op: Dereferencing null pointer "type".
# 773| out:
# 774| if (tconf.expand_types)
# 775|-> --type->recursivity_level;
# 776|
# 777| return printed;
Reported-by: William Cohen <wcohen@redhat.com>
Fixes: f84bf73d54 ("dwarves: Move the fprintf code to a new source file.")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
[ 22%] Building C object CMakeFiles/dwarves.dir/dwarves_fprintf.c.o
/home/acme/git/pahole/dwarves_fprintf.c: In function ‘union__fprintf’:
/home/acme/git/pahole/dwarves_fprintf.c:963:34: error: pointer targets in assignment from ‘int *’ to ‘uint32_t *’ {aka ‘unsigned int *’} differ in signedness [-Werror=pointer-sign]
uconf.cachelinep = &cacheline;
^
cc1: all warnings being treated as errors
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In cf459ca16f ("fprintf: Pretty print struct members that are pointers
to nameless structs") I added some recursive logic that theoretically
may end up doing an overlapping copy as reported by coverity:
Error: OVERLAPPING_COPY: [#def19]
dwarves-1.13/dwarves_fprintf.c:707: assign: Assigning: "name" = "namebfptr".
dwarves-1.13/dwarves_fprintf.c:705: equal: "name" is equal to the address of "namebfptr".
dwarves-1.13/dwarves_fprintf.c:705: overlapping_copy: In the call to function "snprintf", the arguments "name" and "namebfptr" may point to the same object.
# 703| if (tag__is_struct(ptype) || tag__is_union(ptype) ||
# 704| tag__is_enumeration(ptype)) {
# 705|-> snprintf(namebfptr, sizeof(namebfptr), "* %s", name);
# 706| tconf.rel_offset = 1;
# 707|
Look at cf459ca16f to see what this is about, but for now I'm just
checking if this is the case and adding a guard, at some point I'll
address this properly to allow for pointers to pointers to nameless
struct/union/enums.
Reported-by: William Cohen <wcohen@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
As we end up goto'ing to a place where we need that original type and
were having just a NULL pointer, oops, fix it.
Related to:
Error: NULL_RETURNS (CWE-476): [#def18]
dwarves-1.13/dwarves_fprintf.c:727: returned_null: "cu__type" returns "NULL" (checked 54 out of 62 times).
dwarves-1.13/dwarves_fprintf.c:727: var_assigned: Assigning: "type" = "NULL" return value from "cu__type".
dwarves-1.13/dwarves_fprintf.c:686: dereference: Dereferencing "type", which is known to be "NULL".
dwarves-1.13/codiff.c:137: example_assign: Example 1: Assigning: "old_type" = return value from "cu__type(old_cu, old->tag.type)".
dwarves-1.13/codiff.c:141: example_checked: Example 1 (cont.): "old_type" has its value checked in "old_type == NULL".
dwarves-1.13/ctracer.c:356: example_assign: Example 2: Assigning: "type" = return value from "cu__type(cu, tag->type)".
dwarves-1.13/ctracer.c:358: example_checked: Example 2 (cont.): "type" has its value checked in "type == NULL".
dwarves-1.13/dwarves.c:914: example_assign: Example 3: Assigning: "type" = return value from "cu__type(cu, tag->type)".
dwarves-1.13/dwarves.c:916: example_checked: Example 3 (cont.): "type" has its value checked in "type == NULL".
dwarves-1.13/dwarves.c:941: example_assign: Example 4: Assigning: "tag" = return value from "cu__type(cu, var->ip.tag.type)".
dwarves-1.13/dwarves.c:942: example_checked: Example 4 (cont.): "tag" has its value checked in "tag != NULL".
dwarves-1.13/dwarves_emit.c:139: example_assign: Example 5: Assigning: "ptr_type" = return value from "cu__type(cu, type->type)".
dwarves-1.13/dwarves_emit.c:141: example_checked: Example 5 (cont.): "ptr_type" has its value checked in "ptr_type == NULL".
# 684|
# 685| next_type:
# 686|-> switch (type->tag) {
# 687| case DW_TAG_pointer_type:
# 688|
Reported-by: William Cohen <wcohen@redhat.com>
Fixes: 568dae4bd4 ("printf: Fixup printing "const" early with "const void"")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
A cut'n'paste error, noticed by coverity:
Error: PRINTF_ARGS: [#def16]
dwarves-1.13/dwarves_fprintf.c:369: extra_argument: This argument was not used by the format string: "conf->suffix". [Note: The source code implementation of the function has been overridden by a builtin model.]
# 367| */
# 368| if (type->size / 8 != sizeof(int))
# 369|-> printed += fprintf(fp, " __attribute__((__packed__))", conf->suffix);
# 370|
# 371| if (conf->suffix)
Reported-by: William Cohen <wcohen@redhat.com>
Fixes: 28a3bc7add ("fprintf: Support packed enums")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Will Cohen reported this NULL pointer dereference while processing some
object linking with cuda:
#0 0x00007ffff7f91453 in __class__fprintf (class=0x522560, cu=0x40ff80, conf=0x7fffffffa930, fp=0x7ffff7ece780 <_IO_2_1_stdout_>)
at /home/acme/git/pahole/dwarves_fprintf.c:1624
#1 0x00007ffff7f92195 in tag__fprintf (tag=0x522560, cu=0x40ff80, conf=0x7fffffffa930, fp=0x7ffff7ece780 <_IO_2_1_stdout_>)
at /home/acme/git/pahole/dwarves_fprintf.c:1835
#2 0x00007ffff7f90b57 in __class__fprintf (class=0x5224c0, cu=0x40ff80, conf=0x7fffffffaaa0, fp=0x7ffff7ece780 <_IO_2_1_stdout_>)
at /home/acme/git/pahole/dwarves_fprintf.c:1406
#3 0x00007ffff7f92195 in tag__fprintf (tag=0x5224c0, cu=0x40ff80, conf=0x40a200 <conf>, fp=0x7ffff7ece780 <_IO_2_1_stdout_>)
at /home/acme/git/pahole/dwarves_fprintf.c:1835
#4 0x0000000000402d03 in class_formatter (class=0x5224c0, cu=0x40ff80, id=1257) at /home/acme/git/pahole/pahole.c:224
#5 0x0000000000403074 in print_classes (cu=0x40ff80) at /home/acme/git/pahole/pahole.c:319
#6 0x0000000000404bb2 in pahole_stealer (cu=0x40ff80, conf_load=0x40a240 <conf_load>) at /home/acme/git/pahole/pahole.c:1174
#7 0x00007ffff7f9ff73 in finalize_cu (cus=0x40b2b0, cu=0x40ff80, dcu=0x7fffffffacf0, conf=0x40a240 <conf_load>)
at /home/acme/git/pahole/dwarf_loader.c:2227
#8 0x00007ffff7f9ffac in finalize_cu_immediately (cus=0x40b2b0, cu=0x40ff80, dcu=0x7fffffffacf0, conf=0x40a240 <conf_load>)
at /home/acme/git/pahole/dwarf_loader.c:2236
#9 0x00007ffff7fa064c in cus__load_module (cus=0x40b2b0, conf=0x40a240 <conf_load>, mod=0x40d760, dw=0x40e980, elf=0x40b360,
filename=0x7fffffffd5e3 "examples/wcohen/02_Exercise.cuda") at /home/acme/git/pahole/dwarf_loader.c:2389
#10 0x00007ffff7fa0760 in cus__process_dwflmod (dwflmod=0x40d760, userdata=0x40d770, name=0x40d910 "examples/wcohen/02_Exercise.cuda",
base=4194304, arg=0x7fffffffcf10) at /home/acme/git/pahole/dwarf_loader.c:2434
#11 0x00007ffff7f32be1 in dwfl_getmodules () from /lib64/libdw.so.1
#12 0x00007ffff7fa0820 in cus__process_file (cus=0x40b2b0, conf=0x40a240 <conf_load>, fd=3,
filename=0x7fffffffd5e3 "examples/wcohen/02_Exercise.cuda") at /home/acme/git/pahole/dwarf_loader.c:2487
#13 0x00007ffff7fa089c in dwarf__load_file (cus=0x40b2b0, conf=0x40a240 <conf_load>, filename=0x7fffffffd5e3 "examples/wcohen/02_Exercise.cuda")
at /home/acme/git/pahole/dwarf_loader.c:2504
#14 0x00007ffff7f8b0dd in cus__load_file (cus=0x40b2b0, conf=0x40a240 <conf_load>, filename=0x7fffffffd5e3 "examples/wcohen/02_Exercise.cuda")
at /home/acme/git/pahole/dwarves.c:1745
#15 0x00007ffff7f8bc2a in cus__load_files (cus=0x40b2b0, conf=0x40a240 <conf_load>, filenames=0x7fffffffd150)
at /home/acme/git/pahole/dwarves.c:2109
#16 0x0000000000404ff0 in main (argc=2, argv=0x7fffffffd148) at /home/acme/git/pahole/pahole.c:1294
(gdb)
(gdb) p class__name(class, cu)
$6 = 0x5cbb85 "__nv_hdl_helper_trait<__nv_dl_tag<int (*)(int, char**), main, 1u>, void (main(int, char**)::__lambda0::*)(int, double&)const>"
(gdb) p class->type.nr_members
$7 = 0
(gdb) p last
$8 = (struct class_member *) 0x0
(gdb)
So, before checking for bitfield details, first check if there were
members.
Now, if we show all structs/classes in that object file and look for the
above data structure, we find it inside another:
$ pahole examples/wcohen/02_Exercise.cuda
<SNIP>
struct __nv_hdl_helper_trait_outer<false, false, int, Kokkos::View<double**>, Kokkos::View<double*>, Kokkos::View<double*> > {
struct __nv_hdl_helper_trait<__nv_dl_tag<int (*)(int, char**), main, 1u>, void (main(int, char**)::__lambda0::*)(int, double&)const> {
class __nv_hdl_wrapper_t<false, false, __nv_dl_tag<int (*)(int, char**), main, 1u>, void(int, double&), int, Kokkos::View<doubl get<main(int, char**)::__lambda0>(class __lambda0, int, class View<double**>, class View<double*>, class View<double*>);
/* size: 1, cachelines: 0, members: 0 */
/* padding: 1 */
/* last cacheline: 1 bytes */
};
/* size: 1, cachelines: 0, members: 0 */
/* padding: 1 */
/* last cacheline: 1 bytes */
};
<SNIP>
$
Reported-by: William Cohen <wcohen@redhat.com>
Fixes: 13e5b9fc00 ("fprintf: Add unnamed bitfield padding at the end to rebuild original type")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Adding support for suppress_packed and leaving out bitfield paddings at the end
of classes when no alignment info is available accidentally used conf instead
of cconf, where conf can be NULL, which causes cconf to have the default
fprintf config.
Fixes: 986a3b58a8 ("fprintf: Only add bitfield forced paddings when alignment info available")
Fixes: 9a4d719304 ("fprintf: Allow suppressing the inferred __attribute__((__packed__))")
Signed-off-by: Timo Paulssen <timonator@perpetuum-immobile.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When printing const more early we're not considering "const void *", fix
it.
Fixes: ccd67bdb20 ("fprintf: Print "const" for class members more early, in type__fprintf()")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In cf459ca16f ("fprintf: Pretty print struct members that are pointers
to nameless structs") we end up breaking 'pahole --expand_types/-E.
All we need there is to print nameless structs when they are referenced
by a pointer member in a struct, i.e. it is defined only there, inline.
Check if that struct is unnamed before emitting it, keeping that fix
while fixing the regression it caused for --expand_types.
Fixes: cf459ca16f ("fprintf: Pretty print struct members that are pointers to nameless structs")
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
If we have:
inline void process_adjtimex_modes(const struct __kernel_timex * txc, s32 * time_tai)
{
}
And any other struct receiving as a parameter pointers to 'struct
__kerne_timex', then the source file with the above inline, since it
doesn't have any inline expansion, i.e. 'pfunct --compile' generates
just empty function bodies, the types won't be included in the resulting
.o.
Since the original file has the expansions, type types will be there and
thus we will not be able to compare those types, so ask for any 'inline'
to be stripped, so that we keep those types and 'fullcircle' can do its
work.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
