From: Diego Novillo <dnovillo@google.com>
To: Sriraman Tallam <tmsriram@google.com>
Cc: Jason Merrill <jason@redhat.com>, Jan Hubicka <jh@suse.cz>,
Xinliang David Li <davidxl@google.com>,
mark@codesourcery.com, nathan@codesourcery.com,
"H.J. Lu" <hjl.tools@gmail.com>,
Richard Guenther <richard.guenther@gmail.com>,
Uros Bizjak <ubizjak@gmail.com>,
reply@codereview.appspotmail.com,
GCC Patches <gcc-patches@gcc.gnu.org>
Subject: Re: User directed Function Multiversioning via Function Overloading (issue5752064)
Date: Fri, 19 Oct 2012 15:23:00 -0000 [thread overview]
Message-ID: <50816D63.3020908@google.com> (raw)
In-Reply-To: <CAAs8Hmyy9S+Ucg6rTDU3xGMg8JEaa0r3rNjbxhbjuk9H8Sb6wg@mail.gmail.com>
On 2012-10-12 18:19 , Sriraman Tallam wrote:
> When the front-end sees more than one decl for "foo", it calls a target hook to
> determine if they are versions. To prevent duplicate definition errors with other
> versions of "foo", "decls_match" function in cp/decl.c is made to return false
> when 2 decls have are deemed versions by the target. This will make all function
> versions of "foo" to be added to the overload list of "foo".
So, this means that this can only work for C++, right? Or could the
same trickery be done some other way in other FEs?
I see no handling of different FEs. If the user tries to use these
attributes from languages other than C++, we should emit a diagnostic.
> +@deftypefn {Target Hook} tree TARGET_GET_FUNCTION_VERSIONS_DISPATCHER (void *@var{arglist})
> +This hook is used to get the dispatcher function for a set of function
> +versions. The dispatcher function is called to invoke the rignt function
s/rignt/right/
> +version at run-time. @var{arglist} is the vector of function versions
> +that should be considered for dispatch.
> +@end deftypefn
> +
> +@deftypefn {Target Hook} tree TARGET_GENERATE_VERSION_DISPATCHER_BODY (void *@var{arg})
> +This hook is used to generate the dispatcher logic to invoke the right
> +function version at runtime for a given set of function versions.
s/runtime/run-time/
> +@hook TARGET_GET_FUNCTION_VERSIONS_DISPATCHER
> +This hook is used to get the dispatcher function for a set of function
> +versions. The dispatcher function is called to invoke the rignt function
s/rignt/right/
> +version at run-time. @var{arglist} is the vector of function versions
> +that should be considered for dispatch.
> +@end deftypefn
> +
> +@hook TARGET_GENERATE_VERSION_DISPATCHER_BODY
> +This hook is used to generate the dispatcher logic to invoke the right
> +function version at runtime for a given set of function versions.
s/runtime/run-time/
> @@ -288,7 +289,6 @@ mark_store (gimple stmt, tree t, void *data)
> }
> return false;
> }
> -
> /* Create cgraph edges for function calls.
> Also look for functions and variables having addresses taken. */
Don't remove vertical white space, please.
> + {
> + struct cgraph_node *callee = cgraph_get_create_node (decl);
> + /* If a call to a multiversioned function dispatcher is
> + found, generate the body to dispatch the right function
> + at run-time. */
> + if (callee->dispatcher_function)
> + {
> + tree resolver_decl;
> + gcc_assert (callee->function_version.next);
What if callee is the last version in the list? Not sure what you are
trying to check here.
> @@ -8601,9 +8601,22 @@ handle_target_attribute (tree *node, tree name, tr
> warning (OPT_Wattributes, "%qE attribute ignored", name);
> *no_add_attrs = true;
> }
> - else if (! targetm.target_option.valid_attribute_p (*node, name, args,
> - flags))
> - *no_add_attrs = true;
> + else
> + {
> + /* When a target attribute is invalid, it may also be because the
> + target for the compilation unit and the attribute match. For
> + instance, target attribute "xxx" is invalid when -mxxx is used.
> + When used with multiversioning, removing the attribute will lead
> + to duplicate definitions if a default version is provided.
> + So, generate a warning here and remove the attribute. */
> + if (!targetm.target_option.valid_attribute_p (*node, name, args, flags))
> + {
> + warning (OPT_Wattributes,
> + "Invalid target attribute in function %qE, ignored.",
> + *node);
> + *no_add_attrs = true;
If you do this, isn't the compiler going to generate two warning
messages? One for the invalid target attribute, the second for the
duplicate definition.
> @@ -228,6 +228,26 @@ struct GTY(()) cgraph_node {
> struct cgraph_node *prev_sibling_clone;
> struct cgraph_node *clones;
> struct cgraph_node *clone_of;
> +
> + /* Function Multiversioning info. */
> + struct {
> + /* Chains all the semantically identical function versions. The
> + first function in this chain is the default function. */
> + struct cgraph_node *prev;
> + /* If this node is a dispatcher for function versions, this points
> + to the default function version, the first function in the chain. */
> + struct cgraph_node *next;
Why not a VEC of function decls? Seems easier to manage and less size
overhead.
> @@ -3516,8 +3522,8 @@ struct GTY(()) tree_function_decl {
> unsigned looping_const_or_pure_flag : 1;
> unsigned has_debug_args_flag : 1;
> unsigned tm_clone_flag : 1;
> -
> - /* 1 bit left */
> + unsigned versioned_function : 1;
> + /* No bits left. */
You ate the last bit! How rude ;)
> @@ -8132,6 +8176,38 @@ joust (struct z_candidate *cand1, struct z_candida
> && (IS_TYPE_OR_DECL_P (cand1->fn)))
> return 1;
>
> + /* For Candidates of a multi-versioned function, make the version with
s/Candidates/candidates/
> + old_current_function_decl = current_function_decl;
> + push_cfun (DECL_STRUCT_FUNCTION (function_decl));
> + current_function_decl = function_decl;
push_cfun will set current_function_decl for you. No need to keep track
of old_current_function_decl.
> + enum feature_priority
> + {
> + P_ZERO = 0,
> + P_MMX,
> + P_SSE,
> + P_SSE2,
> + P_SSE3,
> + P_SSSE3,
> + P_PROC_SSSE3,
> + P_SSE4_a,
> + P_PROC_SSE4_a,
> + P_SSE4_1,
> + P_SSE4_2,
> + P_PROC_SSE4_2,
> + P_POPCNT,
> + P_AVX,
> + P_AVX2,
> + P_FMA,
> + P_PROC_FMA
> + };
There's no need to have this list dynamically defined, right?
> + }
> + }
> +
> + /* Process feature name. */
> + tok_str = (char *) xmalloc (strlen (attrs_str) + 1);
XNEWVEC(char, strlen (attrs_str) + 1);
> + /* Atleast one more version other than the default. */
s/Atleast/At least/
> + num_versions = VEC_length (tree, fndecls);
> + gcc_assert (num_versions >= 2);
> +
> + function_version_info = (struct _function_version_info *)
> + xmalloc ((num_versions - 1) * sizeof (struct _function_version_info));
Better use VEC() here.
> +
> + /* The first version in the vector is the default decl. */
> + default_decl = VEC_index (tree, fndecls, 0);
> +
> + old_current_function_decl = current_function_decl;
> + push_cfun (DECL_STRUCT_FUNCTION (dispatch_decl));
> + current_function_decl = dispatch_decl;
No need to set current_function_decl.
> +
> + gseq = bb_seq (*empty_bb);
> + /* Function version dispatch is via IFUNC. IFUNC resolvers fire before
> + constructors, so explicity call __builtin_cpu_init here. */
> + ifunc_cpu_init_stmt = gimple_build_call_vec (
> + ix86_builtins [(int) IX86_BUILTIN_CPU_INIT], NULL);
> + gimple_seq_add_stmt (&gseq, ifunc_cpu_init_stmt);
> + gimple_set_bb (ifunc_cpu_init_stmt, *empty_bb);
> + set_bb_seq (*empty_bb, gseq);
> +
> + pop_cfun ();
> + current_function_decl = old_current_function_decl;
Likewise here.
> +/* This function returns true if fn1 and fn2 are versions of the same function.
> + Returns false if only one of the function decls has the target attribute
> + set or if the targets of the function decls are different. This assumes
> + the fn1 and fn2 have the same signature. */
Mention the arguments in capitals.
> + for (i = 0; i < strlen (str); i++)
> + if (str[i] == ',')
> + argnum++;
> +
> + attr_str = (char *)xmalloc (strlen (str) + 1);
XNEWVEC()
> + strcpy (attr_str, str);
> +
> + /* Replace "=,-" with "_". */
> + for (i = 0; i < strlen (attr_str); i++)
> + if (attr_str[i] == '=' || attr_str[i]== '-')
> + attr_str[i] = '_';
> +
> + if (argnum == 1)
> + return attr_str;
> +
> + args = (char **)xmalloc (argnum * sizeof (char *));
VEC()?
> + if (DECL_DECLARED_INLINE_P (decl)
> + && lookup_attribute ("gnu_inline",
> + DECL_ATTRIBUTES (decl)))
> + error_at (DECL_SOURCE_LOCATION (decl),
> + "Function versions cannot be marked as gnu_inline,"
> + " bodies have to be generated\n");
No newline at the end of the error message.
> + sprintf (assembler_name, "%s.%s", orig_name, attr_str);
> + if (dump_file)
> + fprintf (stderr, "Assembler name set to %s for function version %s\n",
> + assembler_name, IDENTIFIER_POINTER (id));
This dumps to stderr instead of dump_file. Also, use the new dumping
facility?
> +/* Return a new name by appending SUFFIX to the DECL name. If
> + make_unique is true, append the full path name. */
Full path name of what?
> +
> +static char *
> +make_name (tree decl, const char *suffix, bool make_unique)
> +{
> + char *global_var_name;
> + int name_len;
> + const char *name;
> + const char *unique_name = NULL;
> +
> + name = IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (decl));
> +
> + /* Get a unique name that can be used globally without any chances
> + of collision at link time. */
> + if (make_unique)
> + unique_name = IDENTIFIER_POINTER (get_file_function_name ("\0"));
> +
> + name_len = strlen (name) + strlen (suffix) + 2;
> +
> + if (make_unique)
> + name_len += strlen (unique_name) + 1;
> + global_var_name = (char *) xmalloc (name_len);
XNEWVEC.
Diego.
next prev parent reply other threads:[~2012-10-19 15:10 UTC|newest]
Thread overview: 93+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-03-07 0:47 Sriraman Tallam
2012-03-07 14:05 ` Richard Guenther
2012-03-07 19:08 ` Sriraman Tallam
2012-03-08 21:37 ` Xinliang David Li
2012-03-08 21:00 ` Xinliang David Li
2012-03-09 20:04 ` Sriraman Tallam
2012-04-27 5:09 ` Sriraman Tallam
2012-04-27 13:39 ` H.J. Lu
2012-04-27 14:35 ` Sriraman Tallam
2012-04-27 14:39 ` H.J. Lu
2012-04-27 14:53 ` Sriraman Tallam
2012-04-27 15:36 ` H.J. Lu
2012-04-27 15:45 ` Sriraman Tallam
2012-05-01 23:51 ` Sriraman Tallam
2012-05-02 0:09 ` H.J. Lu
2012-05-02 2:45 ` Sriraman Tallam
2012-05-02 13:42 ` H.J. Lu
2012-05-02 15:08 ` Sriraman Tallam
2012-05-02 16:06 ` H.J. Lu
2012-05-02 17:44 ` Sriraman Tallam
2012-05-02 18:04 ` H.J. Lu
2012-05-07 16:58 ` Sriraman Tallam
2012-05-09 19:01 ` Sriraman Tallam
2012-05-10 17:55 ` H.J. Lu
2012-05-12 2:04 ` Sriraman Tallam
2012-05-12 13:38 ` H.J. Lu
2012-05-14 18:29 ` Sriraman Tallam
2012-05-26 0:07 ` H.J. Lu
2012-05-26 0:16 ` Sriraman Tallam
2012-05-26 0:27 ` H.J. Lu
2012-05-26 1:54 ` Sriraman Tallam
[not found] ` <CAMe9rOowm9K7r1xnRdRjW5Y4Ay+WxgSsBLTgGvq24z=i42AS+g@mail.gmail.com>
[not found] ` <CAAs8HmzeQigcLQyfkC02u=6gCTLkjLLa_jYmp+b1HEtpMCrYWw@mail.gmail.com>
2012-05-26 5:06 ` H.J. Lu
2012-05-26 22:35 ` Sriraman Tallam
2012-05-26 23:56 ` H.J. Lu
2012-05-27 0:24 ` Sriraman Tallam
2012-05-27 2:06 ` H.J. Lu
2012-05-27 2:23 ` Sriraman Tallam
2012-05-27 2:31 ` H.J. Lu
2012-05-27 19:02 ` Ian Lance Taylor
2012-06-04 19:01 ` Sriraman Tallam
2012-06-04 21:36 ` H.J. Lu
2012-06-04 22:29 ` Sriraman Tallam
2012-06-05 13:56 ` H.J. Lu
2012-06-14 20:35 ` Sriraman Tallam
2012-06-20 1:10 ` Sriraman Tallam
2012-07-06 9:14 ` Richard Guenther
2012-07-06 17:38 ` Sriraman Tallam
2012-07-07 6:06 ` Jason Merrill
2012-07-07 18:38 ` Xinliang David Li
2012-07-08 11:21 ` Jason Merrill
2012-07-09 21:27 ` Xinliang David Li
2012-07-10 9:46 ` Jason Merrill
2012-07-10 16:09 ` Xinliang David Li
[not found] ` <CAAs8HmxHF38ktt6syjWp-MpjiX+6NcXh7_8Xn6iKnAiF2vRymQ@mail.gmail.com>
2012-07-19 20:40 ` Jason Merrill
2012-07-30 19:16 ` Sriraman Tallam
2012-08-25 0:34 ` Sriraman Tallam
2012-09-18 16:29 ` Sriraman Tallam
2012-10-05 17:07 ` Xinliang David Li
2012-10-05 17:44 ` Jason Merrill
2012-10-05 18:14 ` Jason Merrill
2012-10-05 21:58 ` Sriraman Tallam
2012-10-05 22:50 ` Jason Merrill
2012-10-05 23:45 ` Sriraman Tallam
2012-10-05 18:32 ` Jason Merrill
2012-10-11 0:13 ` Sriraman Tallam
2012-10-12 22:41 ` Sriraman Tallam
2012-10-19 15:23 ` Diego Novillo [this message]
2012-10-20 4:29 ` Sriraman Tallam
2012-10-23 21:21 ` Sriraman Tallam
2012-10-26 16:53 ` Jan Hubicka
2012-10-28 4:31 ` Sriraman Tallam
2012-10-29 13:05 ` Jan Hubicka
2012-10-29 17:56 ` Sriraman Tallam
2012-10-30 19:18 ` Jason Merrill
2012-10-31 0:58 ` Sriraman Tallam
[not found] ` <CAAs8Hmw09giv-5_v0irhByTjTJV=kD58rCAD2SAz7M8zrwjBOA@mail.gmail.com>
2012-10-31 14:27 ` Jason Merrill
2012-11-02 2:53 ` Sriraman Tallam
2012-11-06 2:38 ` Sriraman Tallam
2012-11-06 15:52 ` Jason Merrill
2012-11-06 18:17 ` Sriraman Tallam
2012-11-10 1:33 ` Sriraman Tallam
2012-11-12 5:04 ` Jason Merrill
2012-11-13 1:11 ` Sriraman Tallam
2012-11-13 2:39 ` Jason Merrill
2012-11-13 21:57 ` Sriraman Tallam
2012-11-17 22:23 ` H.J. Lu
2012-11-06 22:15 ` Gerald Pfeifer
2012-10-26 14:11 ` Diego Novillo
2012-10-26 16:54 Xinliang David Li
2012-10-26 17:28 ` Sriraman Tallam
2012-11-06 22:17 Dominique Dhumieres
2012-11-07 1:16 ` Gerald Pfeifer
2012-11-07 8:53 ` Dominique Dhumieres
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=50816D63.3020908@google.com \
--to=dnovillo@google.com \
--cc=davidxl@google.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hjl.tools@gmail.com \
--cc=jason@redhat.com \
--cc=jh@suse.cz \
--cc=mark@codesourcery.com \
--cc=nathan@codesourcery.com \
--cc=reply@codereview.appspotmail.com \
--cc=richard.guenther@gmail.com \
--cc=tmsriram@google.com \
--cc=ubizjak@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).