This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[review] Precompute hash value for symbol_set_names
- From: "Tom Tromey (Code Review)" <gerrit at gnutoolchain-gerrit dot osci dot io>
- To: Christian Biesinger <cbiesinger at google dot com>, gdb-patches at sourceware dot org
- Date: Tue, 29 Oct 2019 15:28:03 -0400
- Subject: [review] Precompute hash value for symbol_set_names
- Auto-submitted: auto-generated
- References: <gerrit.1572031795000.I044449e7eb60cffc1c43efd3412f2b485bd9faac@gnutoolchain-gerrit.osci.io>
- Reply-to: gnutoolchain-gerrit at osci dot io
Tom Tromey has posted comments on this change.
Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307
......................................................................
Patch Set 1: Code-Review+1
(4 comments)
This looks good. I found a few nits, see below.
I'll probably put in the threading patches "soon", they've
been few a through rounds of review and have been sitting for
a while now.
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/minsyms.c
File gdb/minsyms.c:
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/minsyms.c@1252
PS1, Line 1252:
1243 | else
1244 | *copyto++ = *copyfrom++;
1245 | }
1246 | *copyto++ = *copyfrom++;
1247 | mcount = copyto - msymbol;
1248 | }
1249 | return (mcount);
1250 | }
1251 |
1252 | struct computed_hash_values {
{ on next line.
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/minsyms.c@1392
PS1, Line 1392:
1370 | #endif
| ...
1383 | /* This will be freed later, by symbol_set_names. */
1384 | char *demangled_name
1385 | = symbol_find_demangled_name (msym, msym->name);
1386 | symbol_set_demangled_name
1387 | (msym, demangled_name,
1388 | &m_objfile->per_bfd->storage_obstack);
1389 | msym->name_set = 1;
1390 |
1391 | size_t idx = msym - msymbols;
1392 | hash_values[idx].mangled_name_hash = htab_hash_string (msym->name);
over-long line? Also, fast_hash
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/symtab.h
File gdb/symtab.h:
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/symtab.h@518
PS1, Line 518:
509 | it. Used for constructs which do not have a mangled name,
510 | e.g. struct tags. Unlike SYMBOL_SET_NAMES, linkage_name must
511 | be terminated and either already on the objfile's obstack or
512 | permanently allocated. */
513 | #define SYMBOL_SET_LINKAGE_NAME(symbol,linkage_name) \
514 | (symbol)->ginfo.name = (linkage_name)
515 |
516 | /* Set the linkage and natural names of a symbol, by demangling
517 | the linkage name. Optionally, HASH can be set to the value
518 | of htab_hash_string (linkage_name) (if nullterminated), to
This has to refer to `fast_hash` now.
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/307/1/gdb/symtab.h@526
PS1, Line 526:
517 | the linkage name. Optionally, HASH can be set to the value
518 | of htab_hash_string (linkage_name) (if nullterminated), to
519 | speed up this function. */
520 | #define SYMBOL_SET_NAMES(symbol,linkage_name,len,copy_name,objfile) \
521 | symbol_set_names (&(symbol)->ginfo, linkage_name, len, copy_name, \
522 | (objfile)->per_bfd)
523 | extern void symbol_set_names (struct general_symbol_info *symbol,
524 | const char *linkage_name, int len, bool copy_name,
525 | struct objfile_per_bfd_storage *per_bfd,
526 | hashval_t hash = 0);
I suppose an optional<hashval_t> would avoid the problem
when the hash value is actually 0
--
Gerrit-Project: binutils-gdb
Gerrit-Branch: master
Gerrit-Change-Id: I044449e7eb60cffc1c43efd3412f2b485bd9faac
Gerrit-Change-Number: 307
Gerrit-PatchSet: 1
Gerrit-Owner: Christian Biesinger <cbiesinger@google.com>
Gerrit-Reviewer: Tom Tromey <tromey@sourceware.org>
Gerrit-Comment-Date: Tue, 29 Oct 2019 19:28:03 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: Yes
Gerrit-MessageType: comment