Skip to contentSkip to conversation

xfstests: fix compile warnings

PATCH In review Dave Chinner 2 participants 18 messages Jan 20, 2010

lore
Share & tools
lore
author
#1 lore raw

PATCH 1/8 xfstests: fix compile warning in doio.c

Use of sigblock is deprecated. Use sigprocmask instead.

Signed-off-by Dave Chinner <[email protected]>
---
 ltp/doio.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
ltp/doio.c +4 -2
index a1c92c9..0fe206d 100644
@@ -350,7 +350,7 @@ char	**argv;
 {
 	int	    	    	i, pid, stat, ex_stat;
 	struct sigaction	sa;
-	int omask;
+	sigset_t		block_mask, old_mask;
 	umask(0);		/* force new file modes to known values */
 #if _CRAYMPP
 	Npes = sysconf(_SC_CRAY_NPES);	/* must do this before parse_cmdline */

@@ -434,7 +434,9 @@ char	**argv;
 		Children[i] = -1;
 	}
 
-	omask = sigblock(sigmask(SIGCLD));
+	sigemptyset(&block_mask);
+	sigaddset(&block_mask, SIGCHLD);
+	sigprocmask(SIG_BLOCK, &block_mask, &old_mask);
 
 	/*
 	 * Fork Nprocs.  This [parent] process is a watchdog, to notify the
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
author
#3 lore raw

PATCH 2/8 xfstests: fix bulkstat related compile warnings.

In reply to Dave Chinner

Use correct types for bulkstat structure to avoid pointer warnings.

Signed-off-by Dave Chinner <[email protected]>
---
 ltp/fsstress.c                      |    4 ++--
 src/bulkstat_unlink_test.c          |    2 +-
 src/bulkstat_unlink_test_modified.c |    2 +-
 src/t_immutable.c                   |    6 +++---
 4 files changed, 7 insertions(+), 7 deletions(-)
ltp/fsstress.c +2 -2
index 6978381..c37cddf 100644
@@ -1550,7 +1550,7 @@ bulkstat_f(int opno, long r)
 {
 	int		count;
 	int		fd;
-	__uint64_t	last;
+	__u64		last;
 	int		nent;
 	xfs_bstat_t	*t;
 	__int64_t	total;

@@ -1583,7 +1583,7 @@ bulkstat1_f(int opno, long r)
 	pathname_t	f;
 	int		fd;
 	int		good;
-	__uint64_t	ino;
+	__u64		ino;
 	struct stat64	s;
 	xfs_bstat_t	t;
 	int		v;
src/bulkstat_unlink_test.c +1 -1
index 17d07c0..afd69ff 100644
@@ -68,7 +68,7 @@ main(int argc, char *argv[])
 		printf("Iteration %d ... (%d files)", k, nfiles);
 
 		memset(&a, 0, sizeof(xfs_fsop_bulkreq_t));
-		a.lastip = &last_inode;
+		a.lastip = (__u64 *)&last_inode;
 		a.icount = nfiles;
 		a.ubuffer = ret;
 		a.ocount = &count;
src/bulkstat_unlink_test_modified.c +1 -1
index 4acc912..3a50d61 100644
@@ -61,7 +61,7 @@ main(int argc, char *argv[])
 	memset(genlist, 0, nfiles * sizeof(__u32));
 	memset(ret, 0, nfiles * sizeof(xfs_bstat_t));
 	memset(&a, 0, sizeof(xfs_fsop_bulkreq_t));
-	a.lastip = &last_inode;
+	a.lastip = (__u64 *)&last_inode;
 	a.icount = nfiles;
 	a.ubuffer = ret;
 	a.ocount = &count;
src/t_immutable.c +3 -3
index 07a36cd..7bb3154 100644
@@ -283,7 +283,7 @@ static int test_immutable(const char *dir)
 
 	  ino = st.st_ino;
 
-	  bulkreq.lastip = &ino;
+	  bulkreq.lastip = (__u64 *)&ino;
 	  bulkreq.icount = 1;
 	  bulkreq.ubuffer = &bstat;
 	  bulkreq.ocount = NULL;

@@ -943,7 +943,7 @@ static int test_append(const char *dir)
 
 	  ino = st.st_ino;
 
-	  bulkreq.lastip = &ino;
+	  bulkreq.lastip = (__u64 *)&ino;
 	  bulkreq.icount = 1;
 	  bulkreq.ubuffer = &bstat;
 	  bulkreq.ocount = NULL;

@@ -1353,7 +1353,7 @@ static int test_append(const char *dir)
 
           ino = st.st_ino;
 
-          bulkreq.lastip = &ino;
+          bulkreq.lastip = (__u64 *)&ino;
           bulkreq.icount = 1;
           bulkreq.ubuffer = &bstat;
           bulkreq.ocount = NULL;
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
author
#5 lore raw

PATCH 3/8 xfstests: Don't use tempnam in growfiles.c

In reply to Dave Chinner

tempnam() generates a compiler warning as a dangerous function. This code doesn't care about security issues with tempnam, so remove it and just manually build the filenames without the randomness of tempnam.

Signed-off-by Dave Chinner <[email protected]>
---
 ltp/growfiles.c |    5 ++---
 1 files changed, 2 insertions(+), 3 deletions(-)
ltp/growfiles.c +2 -3
index a987b5a..52c9079 100644
@@ -1162,10 +1162,9 @@ no whole file checking will be performed!\n", Progname, TagName, (int)getpid());
 	/*
 	 * construct auto filename and insert them into filenames space
 	 */
-		
 	for(ind=0;ind<num_auto_files; ind++, num++) {
-		sprintf((char *)filenames+(num*PATH_MAX), "%s.%d",
-			tempnam(auto_dir, auto_file), ind );
+		sprintf((char *)filenames+(num*PATH_MAX), "%s/%s.%d",
+			auto_dir, auto_file, ind);
 	}
 
 	/*
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 3/8
#6 lore raw

Re: [PATCH 3/8] xfstests: Don't use tempnam in growfiles.c

Quoted 4 lines · Dave Chinner

tempnam() generates a compiler warning as a dangerous function. This code doesn't care about security issues with tempnam, so remove it and just manually build the filenames without the randomness of tempnam.

Looks good,

Reviewed-by Christoph Hellwig <[email protected]>

Btw, I wonder if we should sync back changes to the ltp/* files to LTP..

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

author
#7 lore raw

PATCH 4/8 xfstests: fix printf format warnings in aio-stress.c

In reply to Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 ltp/aio-stress.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
ltp/aio-stress.c +9 -6
index ab9e006..57a2158 100644
@@ -1167,9 +1167,9 @@ restart:
 
     if (t->stage_mb_trans && t->num_files > 0) {
         double seconds = time_since_now(&stage_time);
-	fprintf(stderr, "thread %d %s totals (%.2f MB/s) %.2f MB in %.2fs\n", 
-	        t - global_thread_info, this_stage, t->stage_mb_trans/seconds, 
-		t->stage_mb_trans, seconds);
+	fprintf(stderr, "thread %llu %s totals (%.2f MB/s) %.2f MB in %.2fs\n",
+	        (unsigned long long)(t - global_thread_info), this_stage,
+		t->stage_mb_trans/seconds, t->stage_mb_trans, seconds);
     }
 
     if (num_threads > 1) {

@@ -1449,16 +1449,19 @@ int main(int ac, char **av)
 
     if (file_size < num_contexts * context_offset) {
         fprintf(stderr, "file size %Lu too small for %d contexts\n", 
-	        file_size, num_contexts);
+	        (unsigned long long)file_size, num_contexts);
 	exit(1);
     }
 
-    fprintf(stderr, "file size %LuMB, record size %luKB, depth %d, ios per iteration %d\n", file_size / (1024 * 1024), rec_len / 1024, depth, io_iter);
+    fprintf(stderr, "file size %LuMB, record size %luKB, depth %d, ios per iteration %d\n",
+	    (unsigned long long)file_size / (1024 * 1024),
+	    rec_len / 1024, depth, io_iter);
     fprintf(stderr, "max io_submit %d, buffer alignment set to %luKB\n", 
             max_io_submit, (page_size_mask + 1)/1024);
     fprintf(stderr, "threads %d files %d contexts %d context offset %LuMB verification %s\n", 
             num_threads, num_files, num_contexts, 
-	    context_offset / (1024 * 1024), verify ? "on" : "off");
+	    (unsigned long long)context_offset / (1024 * 1024),
+	    verify ? "on" : "off");
     /* open all the files and do any required setup for them */
     for (i = optind ; i < ac ; i++) {
 	int thread_index;
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 4/8
#8 lore raw

Re: [PATCH 4/8] xfstests: fix printf format warnings in aio-stress.c

Quoted 5 lines
ltp/aio-stress.c +0 -0
index ab9e006..57a2158 100644

ltp/aio-stress.c restart: · line 1172

Quoted 9 lines
ltp/aio-stress.c +3 -3
@@ -1167,9 +1167,9 @@ restart:
 
     if (t->stage_mb_trans && t->num_files > 0) {
         double seconds = time_since_now(&stage_time);
-	fprintf(stderr, "thread %d %s totals (%.2f MB/s) %.2f MB in %.2fs\n", 
-	        t - global_thread_info, this_stage, t->stage_mb_trans/seconds, 
-		t->stage_mb_trans, seconds);
+	fprintf(stderr, "thread %llu %s totals (%.2f MB/s) %.2f MB in %.2fs\n",
+	        (unsigned long long)(t - global_thread_info), this_stage,
+		t->stage_mb_trans/seconds, t->stage_mb_trans, seconds);

Differences between two pointers in C99 is a ptrdiff_t. Probably not worth searching the printf sepcifier for it and just casting it up, so ok.

Reviewed-by Christoph Hellwig <[email protected]>

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

author
#9 lore raw

PATCH 5/8 xfstests: fix compile warnings in iopat.c

In reply to Dave Chinner

Printf format warnings and return value warnings need fixing.

Signed-off-by Dave Chinner <[email protected]>
---
 src/iopat.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
src/iopat.c +2 -1
index c06071f..f34de19 100644
@@ -47,12 +47,13 @@ int main(int argc, char **argv)
    read(fd, &x, 1048576);
    for (i = 0; i < 131072; i++) {
       if (x[i] != i) {
-         printf("error: %d %d %d\n",i,8*i,x[i]);
+         printf("error: %d %d %lld\n", i ,8 * i, (long long)x[i]);
 	 exit(1);
       }
    }
 #endif
 
    close(fd);
+   exit(0);
 }
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 5/8
#10 lore raw

Re: [PATCH 5/8] xfstests: fix compile warnings in iopat.c

Quoted 12 lines · Dave Chinner

Printf format warnings and return value warnings need fixing.

Signed-off-by Dave Chinner <[email protected]>
---
 src/iopat.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
src/iopat.c +0 -0
index c06071f..f34de19 100644

src/iopat.c int main(int argc, char **argv) · line 51

Quoted 6 lines
src/iopat.c +1 -1
@@ -47,12 +47,13 @@ int main(int argc, char **argv)
    read(fd, &x, 1048576);
    for (i = 0; i < 131072; i++) {
       if (x[i] != i) {
-         printf("error: %d %d %d\n",i,8*i,x[i]);
+         printf("error: %d %d %lld\n", i ,8 * i, (long long)x[i]);
 	 exit(1);

Instead of the casting we could also use the C99 macro for the uin64_t format. But either way good enough for the tests,

Reviewed-by Christoph Hellwig <[email protected]>

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

author
#11 lore raw

PATCH 6/8 xfstests: fix printf format warnings in looptest.c

In reply to Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 src/looptest.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)
src/looptest.c +4 -4
index d951f2b..3d99f38 100755
@@ -202,7 +202,7 @@ main(int argc, char *argv[])
         
         if ((flags & FLAG_OPENCLOSE) && (flags & FLAG_SEQUENTIAL)) {
             if (flags & FLAG_VERBOSE)
-                printf("seek %" LL "d\n", seek_to);
+                printf("seek %" LL "d\n", (long long)seek_to);
             if (SEEK(f, seek_to, FILE_BEGIN) < 0)
 	        PERROR("SEEK", GET_LAST_ERROR);
         }

@@ -212,7 +212,7 @@ main(int argc, char *argv[])
             
             if (!(flags & FLAG_SEQUENTIAL)) {
                 if (flags & FLAG_VERBOSE)
-                    printf("seek %" LL "d\n", seek_to);
+                    printf("seek %" LL "d\n", (long long)seek_to);
                 if (SEEK(f, seek_to, FILE_BEGIN) < 0)
 	            PERROR("SEEK", GET_LAST_ERROR);
             }

@@ -233,7 +233,7 @@ main(int argc, char *argv[])
             
             if (!(flags & FLAG_SEQUENTIAL) || (flags & FLAG_WRITE)) {
                 if (flags & FLAG_VERBOSE)
-                    printf("seek %" LL "d\n", seek_to);
+                    printf("seek %" LL "d\n", (long long)seek_to);
                 if (SEEK(f, seek_to, FILE_BEGIN) < 0)
 	            PERROR("SEEK", GET_LAST_ERROR);
             }

@@ -279,7 +279,7 @@ main(int argc, char *argv[])
 	    seek_to += bufsize;
 	    if (flags & FLAG_TRUNCATE) {
                 if (flags & FLAG_VERBOSE)
-                    printf("seek %" LL "d\n", seek_to);
+                    printf("seek %" LL "d\n", (long long)seek_to);
                 if (SEEK(f, seek_to, FILE_BEGIN) < 0)
 	            PERROR("SEEK", GET_LAST_ERROR);
 	    }
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 6/8
#12 lore raw

Re: [PATCH 6/8] xfstests: fix printf format warnings in looptest.c

Quoted 10 lines · Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 src/looptest.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)
src/looptest.c +0 -0
index d951f2b..3d99f38 100755

src/looptest.c main(int argc, char *argv[]) · line 208

Quoted 8 lines
src/looptest.c +1 -1
@@ -202,7 +202,7 @@ main(int argc, char *argv[])
 
         if ((flags & FLAG_OPENCLOSE) && (flags & FLAG_SEQUENTIAL)) {
             if (flags & FLAG_VERBOSE)
-                printf("seek %" LL "d\n", seek_to);
+                printf("seek %" LL "d\n", (long long)seek_to);
             if (SEEK(f, seek_to, FILE_BEGIN) < 0)
 	        PERROR("SEEK", GET_LAST_ERROR);
 	    seek_to += bufsize;

Ok.

Reviewed-by Christoph Hellwig <[email protected]>

What an ugly code in this file, urggg.

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

author
#13 lore raw

PATCH 7/8 xfstests: fix printf warnings in locktest.c

In reply to Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 src/locktest.c |   35 ++++++++++++++++++++---------------
 1 files changed, 20 insertions(+), 15 deletions(-)
src/locktest.c +20 -15
index 7552ace..a108b6f 100644
@@ -718,7 +718,7 @@ send_ctl(void)
 
     if (debug > 1) {
 	fprintf(stderr, "send_ctl: test=%d, command=%d offset=%"LL"d, length=%"LL"d, result=%d, error=%d\n", 
-                ctl.test, ctl.command, ctl.offset, ctl.length,ctl.result, ctl.error);
+                ctl.test, ctl.command, (long long)ctl.offset, (long long)ctl.length,ctl.result, ctl.error);
     }
 
     ctl.test= bswap_uint32(ctl.test);

@@ -741,7 +741,7 @@ send_ctl(void)
         if (nwrite < 0)
             perror("send_ctl: write");
         else
-            fprintf(stderr, "send_ctl[%d]: write() returns %d, not %d as expected\n", 
+            fprintf(stderr, "send_ctl[%d]: write() returns %d, not %lu as expected\n", 
                     ctl.test, nwrite, sizeof(ctl));
         exit(1);
         /*NOTREACHED*/

@@ -756,7 +756,7 @@ void recv_ctl(void)
         if (nread < 0)
             perror("recv_ctl: read");
         else {
-            fprintf(stderr, "recv_ctl[%d]: read() returns %d, not %d as expected\n", 
+            fprintf(stderr, "recv_ctl[%d]: read() returns %d, not %lu as expected\n", 
                     ctl.test, nread, sizeof(ctl));
 	    fprintf(stderr, "socket might has been closed by other locktest\n");
 	} 

@@ -773,7 +773,7 @@ void recv_ctl(void)
 
     if (debug > 1) {
 	fprintf(stderr, "recv_ctl: test=%d, command=%d offset=%"LL"d, length=%"LL"d, result=%d, error=%d\n", 
-                ctl.test, ctl.command, ctl.offset, ctl.length, ctl.result, ctl.error);
+                ctl.test, ctl.command, (long long)ctl.offset, (long long)ctl.length, ctl.result, ctl.error);
     }
 }
 

@@ -1040,14 +1040,16 @@ main(int argc, char *argv[])
 			fail_flag++;
 			/* We have a failure */
 			if(debug)
-			    fprintf(stderr, "Server failure in test %d, while %sing using offset %llu, length %llu - err = %d:%s\n", 
+			    fprintf(stderr, "Server failure in test %d, while %sing using offset %lld, length %lld - err = %d:%s\n", 
 					ctl.test, tests[index][COMMAND]==WRLOCK?"write lock":
 						tests[index][COMMAND]==RDLOCK?"read lock":
 						tests[index][COMMAND]==UNLOCK?"unlock":
 						tests[index][COMMAND]==F_OPEN?"open":"clos", 
-						tests[index][OFFSET], tests[index][LENGTH], saved_errno, strerror(saved_errno));
-			fprintf(stderr, "Server failure in %llu:%s\n",
-					tests[index][TEST_NUM],
+						(long long)tests[index][OFFSET],
+						(long long)tests[index][LENGTH],
+						saved_errno, strerror(saved_errno));
+			fprintf(stderr, "Server failure in %lld:%s\n",
+					(long long)tests[index][TEST_NUM],
 					descriptions[tests[index][TEST_NUM] - 1]);
 		    }
 		}

@@ -1058,12 +1060,13 @@ main(int argc, char *argv[])
 		    end=1;
 		} 
 		if(debug > 1)
-		    fprintf(stderr, "Sending command to client (%d) - %s - %llu:%llu\n", 
+		    fprintf(stderr, "Sending command to client (%d) - %s - %lld:%lld\n", 
 					index, tests[index][COMMAND]==WRLOCK?"write lock":
 					tests[index][COMMAND]==RDLOCK?"read lock":
 					tests[index][COMMAND]==UNLOCK?"unlock": 
 					tests[index][COMMAND]==F_OPEN?"open":"clos", 
-					tests[index][OFFSET], tests[index][LENGTH]);
+					(long long)tests[index][OFFSET],
+					(long long)tests[index][LENGTH]);
 		/* get the client to do something */
 		ctl.index = index;
 		send_ctl();

@@ -1075,14 +1078,15 @@ main(int argc, char *argv[])
 		    if( ctl.result == FAIL ) {
 			fail_flag++;
 			if(debug)
-			    fprintf(stderr, "Client failure in test %d, while %sing using offset %llu, length %llu - err = %d:%s\n",
+			    fprintf(stderr, "Client failure in test %d, while %sing using offset %lld, length %lld - err = %d:%s\n",
 					ctl.test, ctl.command==WRLOCK?"write lock":
 					ctl.command==RDLOCK?"read lock":
 					ctl.command==UNLOCK?"unlock":
 					ctl.command==F_OPEN?"open":"clos",
-					ctl.offset, ctl.length, ctl.error, strerror(ctl.error));
-			fprintf(stderr, "Client failure in %llu:%s\n",
-					tests[index][TEST_NUM],
+					(long long)ctl.offset, (long long)ctl.length,
+					ctl.error, strerror(ctl.error));
+			fprintf(stderr, "Client failure in %lld:%s\n",
+					(long long)tests[index][TEST_NUM],
 					descriptions[tests[index][TEST_NUM] - 1]);
 		    }
 		}

@@ -1144,7 +1148,8 @@ main(int argc, char *argv[])
 	    }
 	    if( result != tests[index][RESULT] ) {
 		if(debug)
-		    fprintf(stderr,"Got %d, wanted %llu\n", result, tests[index][RESULT]);
+		    fprintf(stderr,"Got %d, wanted %lld\n", result,
+					(long long)tests[index][RESULT]);
 		ctl.result = FAIL;
 		ctl.error = saved_errno;
 		fail_count++;
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 7/8
#14 lore raw

Re: [PATCH 7/8] xfstests: fix printf warnings in locktest.c

Quoted 4 lines · Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 src/locktest.c |   35 ++++++++++++++++++++---------------
 1 files changed, 20 insertions(+), 15 deletions(-)

Looks good,

Reviewed-by Christoph Hellwig <[email protected]>

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

author
#15 lore raw

PATCH 8/8 xfstests: don't redefine _GNU_SOURCE in aio tests

In reply to Dave Chinner
Signed-off-by Dave Chinner <[email protected]>
---
 src/aio-dio-regress/aio-dio-subblock-eof-read.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)
src/aio-dio-regress/aio-dio-subblock-eof-read.c +0 -2
index cbcff97..a48841d 100644
@@ -31,8 +31,6 @@
  *  worth of data, and the user issued a 4k read, we want to ensure that
  *  the return value is 300, not 4k.
  */
-
-#define _GNU_SOURCE
 #include <stdio.h>
 #include <stdlib.h>
 #include <libaio.h>
Signature · 6 lines
1.6.5

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
author
#17 lore raw
In reply to Dave Chinner
Quoted 3 lines · Dave Chinner

This series fixes all the compile warnings when building xfstests. It doesn't deal with the dmapi subdirectory but cleans everything else up.

Pushed into xfstests-dev tree.

Signature · 7 lines
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs