All of lore.kernel.org
 help / color / mirror / Atom feed
From: Richard Palethorpe via ltp <ltp@lists.linux.it>
To: ltp@lists.linux.it
Cc: Richard Palethorpe <rpalethorpe@suse.com>
Subject: [LTP] [PATCH 1/3] tools/sparse: Add static check
Date: Tue, 23 Nov 2021 12:43:46 +0000	[thread overview]
Message-ID: <20211123124348.31073-2-rpalethorpe@suse.com> (raw)
In-Reply-To: <20211123124348.31073-1-rpalethorpe@suse.com>

This was adapted from Sparse's inbuilt check_duplicates (-Wdecl). The
original check appears to print a warning whenever a symbol is
non-static, but has no prototype. It appears to work because library
symbols are usually declared first in a header file and then again
with their definition in a source file.

The LTP version also checks for the various library prefixes, but
should otherwise be the same.

Signed-off-by: Richard Palethorpe <rpalethorpe@suse.com>
---
 tools/sparse/sparse-ltp.c | 53 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)

diff --git a/tools/sparse/sparse-ltp.c b/tools/sparse/sparse-ltp.c
index 45874e8eb..73725d191 100644
--- a/tools/sparse/sparse-ltp.c
+++ b/tools/sparse/sparse-ltp.c
@@ -82,6 +82,57 @@ static void do_entrypoint_checks(struct entrypoint *ep)
 	} END_FOR_EACH_PTR(bb);
 }
 
+/* Check for LTP-003 and LTP-004
+ *
+ * Try to find cases where the static keyword was forgotten.
+ */
+static void check_symbol_visibility(struct symbol *sym)
+{
+	struct symbol *next = sym;
+	unsigned long mod = sym->ctype.modifiers;
+	const char *name = show_ident(sym->ident);
+	int has_lib_prefix = !strncmp("tst_", name, 4) ||
+		!strncmp("TST_", name, 4) ||
+		!strncmp("ltp_", name, 4) ||
+		!strncmp("safe_", name, 5);
+
+	if (!(mod & MOD_TOPLEVEL))
+		return;
+
+	if (has_lib_prefix && (mod & MOD_STATIC)) {
+		warning(sym->pos,
+			"LTP-003: Symbol '%s' has the LTP public library prefix, but is static (private).",
+			name);
+		return;
+	}
+
+	if ((mod & MOD_STATIC))
+		return;
+
+	if (tu_kind == LTP_LIB && !has_lib_prefix) {
+		warning(sym->pos,
+			"LTP-003: Symbol '%s' is a public library function, but is missing the 'tst_' prefix",
+			name);
+		return;
+	}
+
+	if (next->same_symbol)
+		return;
+
+	if (sym->ident == &main_ident)
+		return;
+
+	warning(sym->pos,
+		"Symbol '%s' has no prototype or library ('tst_') prefix. Should it be static?",
+		name);
+}
+
+/* AST level checks */
+static void do_symbol_checks(struct symbol *sym)
+{
+	check_symbol_visibility(sym);
+}
+
 /* Compile the AST into a graph of basicblocks */
 static void process_symbols(struct symbol_list *list)
 {
@@ -90,6 +141,8 @@ static void process_symbols(struct symbol_list *list)
 	FOR_EACH_PTR(list, sym) {
 		struct entrypoint *ep;
 
+		do_symbol_checks(sym);
+
 		expand_symbol(sym);
 		ep = linearize_symbol(sym);
 		if (!ep || !ep->entry)
-- 
2.33.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

  reply	other threads:[~2021-11-23 12:44 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-23 12:43 [LTP] [PATCH 0/3] tools/sparse: Add static check Richard Palethorpe via ltp
2021-11-23 12:43 ` Richard Palethorpe via ltp [this message]
2021-11-24  7:12   ` [LTP] [PATCH 1/3] " Petr Vorel
2021-11-30  8:24     ` Richard Palethorpe
2021-11-30 10:19       ` Cyril Hrubis
2021-11-30 10:23         ` Cyril Hrubis
2021-11-30 10:33           ` Richard Palethorpe
2021-11-30 10:51             ` Cyril Hrubis
2021-11-30 11:43             ` Cyril Hrubis
2021-11-24 10:00   ` Richard Palethorpe
2021-11-29 10:16   ` Cyril Hrubis
2021-11-23 12:43 ` [LTP] [PATCH 2/3] doc: Add LTP-003 and LTP-004 static and tst API prefix rules Richard Palethorpe via ltp
2021-11-24  7:18   ` Petr Vorel
2021-11-29 10:17     ` Cyril Hrubis
2021-11-29 10:33       ` Petr Vorel
2021-11-29 10:44   ` Cyril Hrubis
2021-11-23 12:43 ` [LTP] [PATCH 3/3] statx: Add missing static keyword to tcase variable Richard Palethorpe via ltp
2021-11-24  7:19   ` Petr Vorel
2021-11-29 10:16     ` Cyril Hrubis

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=20211123124348.31073-2-rpalethorpe@suse.com \
    --to=ltp@lists.linux.it \
    --cc=rpalethorpe@suse.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.