From: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
To: josh@joshtriplett.org
Cc: Christopher Li <sparse@chrisli.org>,
Christian Borntraeger <borntraeger@de.ibm.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
"Jason J. Herne" <jjherne@linux.vnet.ibm.com>,
Linux-Sparse <linux-sparse@vger.kernel.org>
Subject: Re: [PATCH] Teach sparse about the __COUNTER__ predefined macro
Date: Sat, 24 Jan 2015 00:59:35 +0100 [thread overview]
Message-ID: <20150123235934.GA42931@macpro.local> (raw)
In-Reply-To: <20150123223856.GB1489@cloud>
On Fri, Jan 23, 2015 at 02:38:56PM -0800, josh@joshtriplett.org wrote:
> On Fri, Jan 23, 2015 at 11:23:32PM +0100, Luc Van Oostenryck wrote:
> > On Fri, Jan 23, 2015 at 08:40:17AM -0800, Christopher Li wrote:
> > > I think sparse haven't implement the __COUNTER__ macro. That is why it emit the
> > > error on duplicate entry.
> >
> > The following patch should fix that.
> [...]
> > Subject: [PATCH] Teach sparse about the __COUNTER__ predefined macro.
> >
> > Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
>
> One issue below.
>
> > --- a/pre-process.c
> > +++ b/pre-process.c
> > @@ -181,6 +181,10 @@ static int expand_one_symbol(struct token **list)
> > time(&t);
> > strftime(buffer, 9, "%T", localtime(&t));
> > replace_with_string(token, buffer);
> > + } else if (token->ident == &__COUNTER___ident) {
> > + static int counter;
> > +
> > + replace_with_integer(token, counter++);
>
> This should not use a static counter. GCC and Sparse can run over more
> than one file in one invocation:
>
> $ head test1.c test2.c
> ==> test1.c <==
> __COUNTER__
> __COUNTER__
>
> ==> test2.c <==
> __COUNTER__
> __COUNTER__
>
> $ gcc -E test1.c test2.c
> # 1 "test1.c"
> # 1 "<built-in>"
> # 1 "<command-line>"
> # 1 "/usr/include/stdc-predef.h" 1 3 4
> # 1 "<command-line>" 2
> # 1 "test1.c"
> 0
> 1
> # 1 "test2.c"
> # 1 "<built-in>"
> # 1 "<command-line>"
> # 1 "/usr/include/stdc-predef.h" 1 3 4
> # 1 "<command-line>" 2
> # 1 "test2.c"
> 0
> 1
>
> Notice that the second file starts with __COUNTER__ at 0 again.
>
> The counter *should* keep counting through include files, but needs to
> reset before starting a new top-level compile.
>
> - Josh Triplett
> --
Yes, indeed.
Thanks for bringing to my attention.
Here is a new version of the patch taking care of that.
Luc
---
Subject: [PATCH] Teach sparse about the __COUNTER__ predefined macro.
This macro expands to sequential integral values starting from 0,
and this for each top-level source file.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
ident-list.h | 1 +
pre-process.c | 4 ++++
validation/preprocessor/counter1.c | 12 ++++++++++++
validation/preprocessor/counter2.c | 14 ++++++++++++++
validation/preprocessor/counter2.h | 1 +
validation/preprocessor/counter3.c | 13 +++++++++++++
6 files changed, 45 insertions(+)
create mode 100644 validation/preprocessor/counter1.c
create mode 100644 validation/preprocessor/counter2.c
create mode 100644 validation/preprocessor/counter2.h
create mode 100644 validation/preprocessor/counter3.c
diff --git a/ident-list.h b/ident-list.h
index d5a145f8..b65b667d 100644
--- a/ident-list.h
+++ b/ident-list.h
@@ -108,6 +108,7 @@ __IDENT(__TIME___ident, "__TIME__", 0);
__IDENT(__func___ident, "__func__", 0);
__IDENT(__FUNCTION___ident, "__FUNCTION__", 0);
__IDENT(__PRETTY_FUNCTION___ident, "__PRETTY_FUNCTION__", 0);
+__IDENT(__COUNTER___ident, "__COUNTER__", 0);
/* Sparse commands */
IDENT_RESERVED(__context__);
diff --git a/pre-process.c b/pre-process.c
index 1aa3d2c4..601e0f26 100644
--- a/pre-process.c
+++ b/pre-process.c
@@ -45,6 +45,7 @@
#include "scope.h"
static int false_nesting = 0;
+static int counter_macro; // __COUNTER__ expansion
#define INCLUDEPATHS 300
const char *includepath[INCLUDEPATHS+1] = {
@@ -181,6 +182,8 @@ static int expand_one_symbol(struct token **list)
time(&t);
strftime(buffer, 9, "%T", localtime(&t));
replace_with_string(token, buffer);
+ } else if (token->ident == &__COUNTER___ident) {
+ replace_with_integer(token, counter_macro++);
}
return 1;
}
@@ -1882,6 +1885,7 @@ static void init_preprocessor(void)
sym->normal = 0;
}
+ counter_macro = 0;
}
static void handle_preprocessor_line(struct stream *stream, struct token **line, struct token *start)
diff --git a/validation/preprocessor/counter1.c b/validation/preprocessor/counter1.c
new file mode 100644
index 00000000..98187ee6
--- /dev/null
+++ b/validation/preprocessor/counter1.c
@@ -0,0 +1,12 @@
+__COUNTER__
+__COUNTER__
+/*
+ * check-name: __COUNTER__ #1
+ * check-command: sparse -E $file
+ *
+ * check-output-start
+
+0
+1
+ * check-output-end
+ */
diff --git a/validation/preprocessor/counter2.c b/validation/preprocessor/counter2.c
new file mode 100644
index 00000000..9883b682
--- /dev/null
+++ b/validation/preprocessor/counter2.c
@@ -0,0 +1,14 @@
+__FILE__ __COUNTER__
+#include <counter2.h>
+__FILE__ __COUNTER__
+/*
+ * check-name: __COUNTER__ #2
+ * check-command: sparse -Ipreprocessor -E $file
+ *
+ * check-output-start
+
+"preprocessor/counter2.c" 0
+"preprocessor/counter2.h" 1
+"preprocessor/counter2.c" 2
+ * check-output-end
+ */
diff --git a/validation/preprocessor/counter2.h b/validation/preprocessor/counter2.h
new file mode 100644
index 00000000..447b70ab
--- /dev/null
+++ b/validation/preprocessor/counter2.h
@@ -0,0 +1 @@
+__FILE__ __COUNTER__
diff --git a/validation/preprocessor/counter3.c b/validation/preprocessor/counter3.c
new file mode 100644
index 00000000..1449b2d1
--- /dev/null
+++ b/validation/preprocessor/counter3.c
@@ -0,0 +1,13 @@
+/*
+ * check-name: __COUNTER__ #3
+ * check-command: sparse -Ipreprocessor -E preprocessor/counter1.c preprocessor/counter2.c
+ *
+ * check-output-start
+
+0
+1
+"preprocessor/counter2.c" 0
+"preprocessor/counter2.h" 1
+"preprocessor/counter2.c" 2
+ * check-output-end
+ */
next prev parent reply other threads:[~2015-01-23 23:59 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-01-22 20:31 sparse: new feature " multiple initializers" has false positives on MODULE_ALIAS Christian Borntraeger
2015-01-23 16:40 ` Christopher Li
2015-01-23 22:23 ` [PATCH] Teach sparse about the __COUNTER__ predefined macro Luc Van Oostenryck
2015-01-23 22:28 ` Sam Ravnborg
2015-01-23 22:38 ` josh
2015-01-23 23:59 ` Luc Van Oostenryck [this message]
2015-01-24 1:29 ` Josh Triplett
2015-01-24 11:27 ` Luc Van Oostenryck
2015-01-24 20:19 ` Josh Triplett
2015-01-24 20:39 ` Luc Van Oostenryck
2015-01-25 20:12 ` Christian Borntraeger
2015-01-28 10:08 ` Christian Borntraeger
2015-02-02 5:17 ` Christopher Li
2015-02-04 2:38 ` Luc Van Oostenryck
2015-02-12 20:16 ` Christopher Li
2015-02-04 2:46 ` [PATCH 1/3] test-suite: add support for tests case involving several input files Luc Van Oostenryck
2015-02-06 15:02 ` Christopher Li
2015-02-04 2:49 ` [PATCH 2/3] test-suite: allow filename expansion of the input sections Luc Van Oostenryck
2015-02-04 2:51 ` [PATCH 3/3] test-suite: consolidate tests that require include files into single test files Luc Van Oostenryck
2015-02-04 3:11 ` [PATCH 2/3] test-suite: allow filename expansion of the input sections Luc Van Oostenryck
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=20150123235934.GA42931@macpro.local \
--to=luc.vanoostenryck@gmail.com \
--cc=borntraeger@de.ibm.com \
--cc=jjherne@linux.vnet.ibm.com \
--cc=josh@joshtriplett.org \
--cc=linux-sparse@vger.kernel.org \
--cc=sparse@chrisli.org \
--cc=torvalds@linux-foundation.org \
/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 NNTP newsgroup(s).