[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505 Sergei Trofimovich changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Assignee|unassigned at gcc dot gnu.org |slyfox at gcc dot gnu.org Resolution|--- |FIXED --- Comment #7 from Sergei Trofimovich --- I can confirm --with-build-config=bootstrap-asan fails the same way before the fix and passes successfully after the fix. Let's declare it FIXED.
[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505 --- Comment #6 from CVS Commits --- The master branch has been updated by Sergei Trofimovich : https://gcc.gnu.org/g:7525707c5f3edb46958c4fdfbe30de5ddfa8923a commit r14-4327-g7525707c5f3edb46958c4fdfbe30de5ddfa8923a Author: Sergei Trofimovich Date: Thu Sep 28 20:20:31 2023 +0100 ggc: do not wipe out unrelated data via gt_ggc_rtab [PR111505] There are 3 GC root tables: gt_ggc_rtab gt_ggc_deletable_rtab gt_pch_scalar_rtab `deletable` and `scalar` tables are both simple: each element always contains a pointer to the beginning of the object and it's size is the full object. `rtab` is different: it's `base` is a pointer in the middle of the struct and `stride` points to the next GC pointer in the array. Before the change there were 2 problems: 1. We memset()ed not just pointers but data around them. 2. We wen out of bounds of the last object described by gt_ggc_rtab and triggered bootstrap failures in profile and asan bootstraps. After the change we handle only pointers themselves like the rest of ggc-common.cc code. gcc/ PR middle-end/111505 * ggc-common.cc (ggc_zero_out_root_pointers, ggc_common_finalize): Add new helper. Use helper instead of memset() to wipe out pointers.
[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505 --- Comment #5 from Sergei Trofimovich --- (In reply to Sergei Trofimovich from comment #4) > Is it a ggc_common_finalize() bug in assuming that `base` does not point to > the bbeginning of a struct? > Or a `gt_ggc_r_gt_cp_tree_h` bug that it does not point to the beginning of > the struct? > > Also while at it: what should `ggc_common_finalize ()` ideally do to > `ovl_op_info`? memset() all of it? Or only pointers? I think it has > information only to do latter but it does it in a very strange way. Proposed fix as https://gcc.gnu.org/pipermail/gcc-patches/2023-September/631634.html . That fixes `profiledbootstrap` for me. Having stared a bit more at gcc/ggc-common.cc the answer is: it's complicated. There are 3 roots: gt_ggc_rtab gt_ggc_deletable_rtab gt_pch_scalar_rtab Last two are simple: these are single-element complete values without complexities. But gt_ggc_rtab is different: it encodes pointer locations within structs (of non-pointers). GCH code carefully extracts and restores those pointers. An example: for (rt = gt_ggc_rtab; *rt; rt++) for (rti = *rt; rti->base != NULL; rti++) for (i = 0; i < rti->nelt; i++) (*rti->pchw)(*(void **)((char *)rti->base + rti->stride * i)); finalize() should do the same.
[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505
--- Comment #4 from Sergei Trofimovich ---
In https://gcc.gnu.org/PR111629#c0 profiled bootstrap fales for a similar
reason.
There ggc_common_finalize() memset()s unexpected memory location
#1 0x01933651 in ggc_common_finalize () at
/home/slyfox/dev/git/gcc/gcc/ggc-common.cc:1312
1312 memset (rti->base, 0, rti->stride * rti->nelt);
(gdb) list
1307for (rti = *rt; rti->base != NULL; rti++)
1308 memset (rti->base, 0, rti->stride * rti->nelt);
1309
1310 for (rt = gt_ggc_rtab; *rt; rt++)
1311for (rti = *rt; rti->base != NULL; rti++)
1312 memset (rti->base, 0, rti->stride * rti->nelt);
for this global:
struct GTY(()) ovl_op_info_t {
/* The IDENTIFIER_NODE for the operator. */
tree identifier;
/* The name of the operator. */
const char *name;
/* The mangled name of the operator. */
const char *mangled_name;
/* The (regular) tree code. */
enum tree_code tree_code : 16;
/* The (compressed) operator code. */
enum ovl_op_code ovl_op_code : 8;
/* The ovl_op_flags of the operator */
unsigned flags : 8;
};
/* Overloaded operator info indexed by ass_op_p & ovl_op_code. */
extern GTY(()) ovl_op_info_t ovl_op_info[2][OVL_OP_MAX];
Generated tables:
/* Structures for the easy way to mark roots.
In an array, terminated by having base == NULL. */
struct ggc_root_tab {
void *base;
size_t nelt;
size_t stride;
gt_pointer_walker cb;
gt_pointer_walker pchw;
};
EXPORTED_CONST struct ggc_root_tab gt_ggc_r_gt_cp_tree_h[] = {
...
{ // 4
_op_info[0][0].identifier,
1 * (2) * (OVL_OP_MAX),
sizeof (ovl_op_info[0][0]),
_ggc_mx_tree_node,
_pch_nx_tree_node
},
{ // 5
_op_info[0][0].name,
1 * (2) * (OVL_OP_MAX),
sizeof (ovl_op_info[0][0]),
(gt_pointer_walker) _ggc_m_S,
(gt_pointer_walker) _pch_n_S
},
{ //6
_op_info[0][0].mangled_name,
1 * (2) * (OVL_OP_MAX),
sizeof (ovl_op_info[0][0]),
(gt_pointer_walker) _ggc_m_S,
(gt_pointer_walker) _pch_n_S
},
Is it a ggc_common_finalize() bug in assuming that `base` does not point to the
bbeginning of a struct?
Or a `gt_ggc_r_gt_cp_tree_h` bug that it does not point to the beginning of the
struct?
Also while at it: what should `ggc_common_finalize ()` ideally do to
`ovl_op_info`? memset() all of it? Or only pointers? I think it has information
only to do latter but it does it in a very strange way.
[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505 Andrew Pinski changed: What|Removed |Added CC||slyfox at gcc dot gnu.org --- Comment #3 from Andrew Pinski --- *** Bug 111629 has been marked as a duplicate of this bug. ***
[Bug middle-end/111505] [14 Regression] Asan (address-sanitizer) bootstrap fails since r14-4003-geaa8e8541349df
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111505 Richard Biener changed: What|Removed |Added Priority|P3 |P1 Summary|Asan (address-sanitizer)|[14 Regression] Asan |bootstrap fails since |(address-sanitizer) |r14-4003-geaa8e8541349df|bootstrap fails since ||r14-4003-geaa8e8541349df --- Comment #2 from Richard Biener --- Quite odd indeed.
