Skip to contentSkip to conversation

tests/cpu-policy: fix format-overflow warning by null terminating strings

PATCH In review [email protected] 6 participants 10 messages Jul 25, 2019 – Aug 2, 2019

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tests/cpu-policy: fix format-overflow warning by null terminating strings

From: Christopher Clark <[email protected]>

gcc 9.1.0 reports:

| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~

so increase the string array size by one character for the null string terminator and add another single char to the struct within the same union to balance it.

---
 tools/tests/cpu-policy/test-cpu-policy.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/tests/cpu-policy/test-cpu-policy.c +2 -1
index ca3b8dd45f..c91408a93a 100644
@@ -35,9 +35,10 @@ static void test_vendor_identification(void)
 {
     static const struct test {
         union {
-            char ident[12];
+            char ident[13];
             struct {
                 uint32_t b, d, c;
+                char null_terminator;
             };
         };
         unsigned int vendor;
Signature · 7 lines
2.17.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
#1 lore raw
Quoted 17 lines · On Wed, Jul 24, 2019 at 05:53:26PM -0700, [email protected] wrote:

From: Christopher Clark <[email protected]>

gcc 9.1.0 reports:

| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~
Quoted 5 lines

so increase the string array size by one character for the null string terminator and add another single char to the struct within the same union to balance it.

Thanks! Just two nits below.

Quoted 9 lines
---
 tools/tests/cpu-policy/test-cpu-policy.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/tests/cpu-policy/test-cpu-policy.c +0 -0
index ca3b8dd45f..c91408a93a 100644

tools/tests/cpu-policy/test-cpu-policy.c static void test_vendor_identification(void) · line 41

Quoted 8 lines
tools/tests/cpu-policy/test-cpu-policy.c +2 -1
@@ -35,9 +35,10 @@ static void test_vendor_identification(void)
 {
     static const struct test {
         union {
-            char ident[12];
+            char ident[13];
             struct {
                 uint32_t b, d, c;
+                char null_terminator;

Do you really need this field here, AFAICT it's unused. The compiler will add the padding here automatically to match the size of the other field of the union.

Also, since ident is null terminated now you can remove the .12 from the format string.

Thanks, Roger.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#2 lore raw
Quoted 15 lines · Roger Pau Monné
Quoted 13 lines · On Wed, Jul 24, 2019 at 05:53:26PM -0700, [email protected] wrote:

From: Christopher Clark <[email protected]>

tools/tests/cpu-policy/test-cpu-policy.c +1 -1
index ca3b8dd45f..c91408a93a 100644
@@ -35,9 +35,10 @@ static void test_vendor_identification(void)
 {
     static const struct test {
         union {
-            char ident[12];
+            char ident[13];
Quoted 9 lines

tools/tests/cpu-policy/test-cpu-policy.c static void test_vendor_identification(void) · line 41

Quoted 3 lines
tools/tests/cpu-policy/test-cpu-policy.c +1 -0
             struct {
                 uint32_t b, d, c;
+                char null_terminator;

Do you really need this field here, AFAICT it's unused. The compiler will add the padding here automatically to match the size of the other
field of the union.

Yep, I run into this myself (and sent a mail about it yesterday), and, FWIW, I confirm that just growing the array to 13 both builds, with GCC9, and works. :-)

Regards

Signature · 6 lines
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
-------------------------------------------------------------------
<<This happens because _I_ choose it to happen!>> (Raistlin Majere)

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#3 lore raw
In reply to [email protected]
Quoted 14 lines · On 25.07.2019 02:53, [email protected] wrote:

gcc 9.1.0 reports:

| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~

I must be missing something here: %.12s says that no more than 12 bytes are to be read from the string. There's nul terminator required. This is what the standard says

"Characters from the array are written up to (but not including)
the terminating null character. If the precision is specified, no more than that many bytes are written. If the precision is not specified or is greater than the size of the array, the array shall contain a null character."

For the moment it looks to me as if the compiler was wrong to complain.

Jan
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#4 lore raw
Quoted 19 lines · Jan Beulich
Quoted 14 lines · On 25.07.2019 02:53, [email protected] wrote:

gcc 9.1.0 reports:

| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~

I must be missing something here: %.12s says that no more than 12 bytes are to be read from the string. There's nul terminator required. This is what the standard says

Quoted 8 lines

"Characters from the array are written up to (but not including)
the terminating null character. If the precision is specified, no more than that many bytes are written. If the precision is not specified or is greater than the size of the array, the array shall contain a null character."

For the moment it looks to me as if the compiler was wrong to complain.

That is my assessment as well, but I haven't had time to file a GCC bug yet.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#5 lore raw
In reply to [email protected]
Quoted 4 lines · On 25.07.2019 02:53, [email protected] wrote:

From: Christopher Clark <[email protected]>

gcc 9.1.0 reports:

Quoted 12 lines
| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~

In order to possibly create a bug report against gcc I've tried this:

#include <stdio.h>

struct s {
	char ac[12];
	int i;
};
void test(const struct s*ps) {
	printf("'%.12s'\n", ps->ac);
}

There's no warning here. Could you check whether the compiler warns on that simple test for you? If it does - are we talking about plain upstream 9.1.0 (in which case I'd be really puzzled by the difference in behavior)?

Jan
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#6 lore raw
Quoted 26 lines · Jan Beulich
Quoted 4 lines · On 25.07.2019 02:53, [email protected] wrote:

From: Christopher Clark <[email protected]>

gcc 9.1.0 reports:

Quoted 12 lines
| test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| test-cpu-policy.c:20:12: note: in definition of macro 'fail'
|    20 |     printf(fmt, ##__VA_ARGS__);                 \
|       |            ^~~
| test-cpu-policy.c:64:27: note: format string is defined here
|    64 |             fail("  Test '%.12s', expected vendor %u, got %u\n",
|       |                           ^~~~~
| test-cpu-policy.c:44:7: note: referenced argument declared here
|    44 |     } tests[] = {
|       |       ^~~~~

In order to possibly create a bug report against gcc I've tried this:

#include <stdio.h>

struct s {
	char ac[12];
	int i;
};
Quoted 8 lines
void test(const struct s*ps) {
	printf("'%.12s'\n", ps->ac);
}

There's no warning here. Could you check whether the compiler warns on that simple test for you? If it does - are we talking about plain upstream 9.1.0 (in which case I'd be really puzzled by the difference in behavior)?

CC Dario who also reported this build failure.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#7 lore raw
Quoted 36 lines · Andrew Cooper
Quoted 35 lines · Jan Beulich
Quoted 20 lines · On 25.07.2019 02:53, [email protected] wrote:

From: Christopher Clark <[email protected]>

gcc 9.1.0 reports:

Quoted 16 lines

test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
64 | fail(" Test '%.12s', expected vendor %u,
got %u\n",

      |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

~~~~~~~~
test-cpu-policy.c:20:12: note: in definition of macro 'fail'
20 | printf(fmt, ##__VA_ARGS__); \
| ^~~
test-cpu-policy.c:64:27: note: format string is defined here
64 | fail(" Test '%.12s', expected vendor %u,
got %u\n",
| ^~~~~
test-cpu-policy.c:44:7: note: referenced argument declared here

   44 |     } tests[] = {
      |       ^~~~~

In order to possibly create a bug report against gcc I've tried this:

#include <stdio.h>

struct s {
	char ac[12];
	int i;
};
void test(const struct s*ps) {
	printf("'%.12s'\n", ps->ac);
}
Quoted 9 lines
Quoted 6 lines

There's no warning here. Could you check whether the compiler warns on
that simple test for you? If it does - are we talking about plain upstream 9.1.0 (in which case I'd be really puzzled by the difference
in behavior)?

CC Dario who also reported this build failure.

Yep, this thread was on my radar already. But thanks! :-P

So, I'm using the following gcc (stock opensuse Tumbleweed provided one):

dario@Palanthas:~/src/xen/xen.git> gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib64/gcc/x86_64-suse-linux/9/lto-wrapper
OFFLOAD_TARGET_NAMES=hsa:nvptx-none
Target: x86_64-suse-linux
Configured with: ../configure --prefix=/usr --infodir=/usr/share/info --mandir=/usr/share/man --libdir=/usr/lib64 --libexecdir=/usr/lib64 --enable-languages=c,c++,objc,fortran,obj-c++,ada,go,d --enable-offload-targets=hsa,nvptx-none=/usr/nvptx-none, --without-cuda-driver --disable-werror --with-gxx-include-dir=/usr/include/c++/9 --enable-ssp --disable-libssp --disable-libvtv --disable-cet --disable-libcc1 --enable-plugin --with-bugurl=https://bugs.opensuse.org/ --with-pkgversion='SUSE Linux' --with-slibdir=/lib64 --with-system-zlib --enable-libstdcxx-allocator=new --disable-libstdcxx-pch --enable-libphobos --enable-version-specific-runtime-libs --with-gcc-major-version-only --enable-linker-build-id --enable-linux-futex --enable-gnu-indirect-function --program-suffix=-9 --without-system-libunwind --enable-multilib --with-arch-32=x86-64 --with-tune=generic --with-build-config=bootstrap-lto-lean --enable-link-mutex --build=x86_64-suse-linux --host=x86_64-suse-linux Thread model: posix
gcc version 9.1.1 20190703 [gcc-9-branch revision 273008] (SUSE Linux)

Jan's example above, seem to compile **without any warnings** for me as well. If I add a main(), I can even get the code above to print the content of the array.

And yet, building the tools without a patch like Christoper's one (which was also what I was using locally, and raised to Andy), I get:

make[6]: Entering directory '/build/tools/tests/cpu-policy'
gcc -m64 -DBUILD_ID -fno-strict-aliasing -std=gnu99 -Wall -Wstrict-prototypes -Wdeclaration-after-statement -Wno-unused-but-set-variable -Wno-unused-local-typedefs -g3 -O0 -fno-omit-frame-pointer -D__XEN_INTERFACE_VERSION__=__XEN_LATEST_INTERFACE_VERSION__ -MMD -MF .test-cpu-policy.o.d -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -Werror -I/build/tools/tests/cpu-policy/../../../tools/include -D__XEN_TOOLS__ -O3 -c -o test-cpu-policy.o test-cpu-policy.c
test-cpu-policy.c: In function 'main':
test-cpu-policy.c:64:18: error: '%.12s' directive argument is not a nul-terminated string [-Werror=format-overflow=]
64 | fail(" Test '%.12s', expected vendor %u, got %u\n",

      |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

test-cpu-policy.c:20:12: note: in definition of macro 'fail'
20 | printf(fmt, ##__VA_ARGS__); \
| ^~~
test-cpu-policy.c:64:27: note: format string is defined here
64 | fail(" Test '%.12s', expected vendor %u, got %u\n",
| ^~~~~
test-cpu-policy.c:44:7: note: referenced argument declared here

   44 |     } tests[] = {
      |       ^~~~~

I'm happy to do more tests.

Regards

Signature · 6 lines
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
-------------------------------------------------------------------
<<This happens because _I_ choose it to happen!>> (Raistlin Majere)

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#8 lore raw
Quoted 6 lines · Dario Faggioli

Jan's example above, seem to compile **without any warnings** for me as well. If I add a main(), I can even get the code above to print the content of the array.

And yet, building the tools without a patch like Christoper's one (which was also what I was using locally, and raised to Andy), I get:

Sure - I'm aware that we're using a (potentially heavily) patched gcc. Hence my question to Christopher whether he's using a plain one.

Jan
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#9 lore raw
Quoted 7 lines · Jan Beulich
Quoted 6 lines · Dario Faggioli

Jan's example above, seem to compile **without any warnings** for me as well. If I add a main(), I can even get the code above to print the content of the array.

Quoted 5 lines
Quoted 2 lines

And yet, building the tools without a patch like Christoper's one (which was also what I was using locally, and raised to Andy), I get:

Sure - I'm aware that we're using a (potentially heavily) patched gcc. Hence my question to Christopher whether he's using a plain one.

The Xen test-cpu-policy.c build error and workaround reported by Christopher can be reproduced by building Xen staging against upstream OpenEmbedded/Yocto meta-virtualization master and GCC. I've tested on Devuan Ascii with "oe-build-xen x86-64 staging master":

http://github.com/dozylynx/oe-build-xen

If there is an upstream GCC compiler issue, Xen will not be the only affected project. We can expect new data as other projects start building with recent GCC versions. In the meantime, the workaround can unblock Xen builds.

Rich

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel