GCC 8 format-truncation error #126

afaerber · 2018-06-17T11:28:50Z

Building v3.1.0 on openSUSE Tumbleweed with its GCC 8.1.1 leads to the following build error:

[   67s] /home/abuild/rpmbuild/BUILD/optee_client-3.1.0/libteec/src/teec_trace.c: In function '_dprintf':
[   67s] /home/abuild/rpmbuild/BUILD/optee_client-3.1.0/libteec/src/teec_trace.c:110:24: error: '%s' directive output may be truncated writing up to 255 bytes into a region of size 246 [-Werror=format-truncation=]
[   67s]      "%s [%d] %s:%s:%d: %s",
[   67s]                         ^~
[   67s] /home/abuild/rpmbuild/BUILD/optee_client-3.1.0/libteec/src/teec_trace.c:112:11:
[   67s]      line, raw);
[   67s]            ~~~           
[   67s] In file included from /usr/include/stdio.h:862,
[   67s]                  from /home/abuild/rpmbuild/BUILD/optee_client-3.1.0/libteec/src/teec_trace.c:27:
[   67s] /usr/include/bits/stdio2.h:64:10: note: '__builtin___snprintf_chk' output 11 or more bytes (assuming 266) into a destination of size 256
[   67s]    return __builtin___snprintf_chk (__s, __n, __USE_FORTIFY_LEVEL - 1,
[   67s]           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[   67s]         __bos (__s), __fmt, __va_arg_pack ());
[   67s]         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[   67s] cc1: all warnings being treated as errors
[   67s] make[2]: *** [libteec/CMakeFiles/teec.dir/build.make:79: libteec/CMakeFiles/teec.dir/src/teec_trace.c.o] Error 1

The text was updated successfully, but these errors were encountered:

afaerber · 2018-06-17T11:35:21Z

@scarabeusiv

simonqhughes · 2018-06-18T09:32:12Z

Hi

This is also an issue for us on our development project, which has broken our CI an resulted in us pinning our repo manifests to get over the problem. Any escalation on priority to fix this would be much appreciated:

> ===================================================================
> 02:48:15 ERROR: Logfile of failure stored in: /work/machine-imx7s-warp-xxx/xxx-manifest/build-xxx/tmp-xxx-glibc/work/cortexa7hf-neon-oe-linux-gnueabi/optee-client/2.6.0+gitAUTOINC+73b4e490a8-r0/temp/log.do_compile.12265
> 02:48:15 Log data follows:
> 02:48:15 | DEBUG: Executing shell function do_compile
> 02:48:15 | NOTE: make -j 24
> 02:48:15 | Building libteec.so
> 02:48:15 |   CC      src/tee_client_api.c
> 02:48:15 |   CC      src/teec_trace.c
> 02:48:15 | src/teec_trace.c: In function '_dprintf':
> 02:48:15 | src/teec_trace.c:110:24: error: '%s' directive output may be truncated writing up to 255 bytes into a region of size 246 [-Werror=format-truncation=]
> 02:48:15 |      "%s [%d] %s:%s:%d: %s",
> 02:48:15 |                         ^~
> 02:48:15 | src/teec_trace.c:112:11:
> 02:48:15 |      line, raw);
> 02:48:15 |            ~~~
> 02:48:15 | src/teec_trace.c:109:3: note: 'snprintf' output 11 or more bytes (assuming 266) into a destination of size 256
> 02:48:15 |    snprintf(prefixed, MAX_PRINT_SIZE,
> 02:48:15 |    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 02:48:15 |      "%s [%d] %s:%s:%d: %s",
> 02:48:15 |      ~~~~~~~~~~~~~~~~~~~~~~~
> 02:48:15 |      trace_level_strings[level], thread_id, prefix, func,
> 02:48:15 |      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 02:48:15 |      line, raw);
> 02:48:15 |      ~~~~~~~~~~
> 02:48:15 | cc1: all warnings being treated as errors
> 02:48:15 | Makefile:52: recipe for target '/work/machine-imx7s-warp-xxx/xxx-manifest/build-xxx/tmp-xxx-glibc/work/cortexa7hf-neon-oe-linux-gnueabi/optee-client/2.6.0+gitAUTOINC+73b4e490a8-r0/git/libteec/../out/libteec/teec_trace.o' failed
> 02:48:15 | make[1]: *** [/work/machine-imx7s-warp-xxx/xxx-manifest/build-xxx/tmp-xxx-glibc/work/cortexa7hf-neon-oe-linux-gnueabi/optee-client/2.6.0+gitAUTOINC+73b4e490a8-r0/git/libteec/../out/libteec/teec_trace.o] Error 1
> 02:48:15 | make[1]: *** Waiting for unfinished jobs....
> 02:48:15 | Makefile:31: recipe for target 'build-libteec' failed
> 02:48:15 | make: *** [build-libteec] Error 2
> 02:48:15 | ERROR: oe_runmake failed
> 02:48:15 | WARNING: exit code 1 from a shell command.
> 02:48:15 | ERROR: Function failed: do_compile (log file is located at /work/machine-imx7s-warp-xxx/xxx-manifest/build-xxx/tmp-xxx-glibc/work/cortexa7hf-neon-oe-linux-gnueabi/optee-client/2.6.0+gitAUTOINC+73b4e490a8-r0/temp/log.do_compile.12265)
> 02:48:15 NOTE: recipe optee-client-2.6.0+gitAUTOINC+73b4e490a8-r0: task do_compile: Failed
> 02:48:15 NOTE: recipe alsa-state-0.2.0-r5: task do_install: Succeeded
> 02:48:15 ERROR: Task (/work/machine-imx7s-warp-xxx/xxx-manifest/build-xxx/conf/../../layers/meta-linaro/meta-optee/recipes-security/optee/optee-client.bb:do_compile) failed with exit code '1'
> 02:48:15 NOTE: recipe modutils-initscripts-1.0-r7: task do_install: Started
> ===================================================================
>  
> between
> openembedded-core master commit bad: 8ab5b439ea82ac775494a0ce7a6f3615b61c94be
> openembedded-core master commit good:23f15c63777020f5d43b070a1eb2bcf246c19ff8
>  
>  
>  
> Here is the diff between xxx-manifest pinned-manifest.xml's between <bad build and >good build
>  
> simhug01@xxx-test:/mnt/data/default_20180613_122141/xxx-manifest$ cat ../manifest_xml_changes.txt 
> 20,21c20,21
> <   <project name="openembedded/meta-openembedded" path="layers/meta-openembedded" remote="github" revision="d9e257abbe16b9d30171493fa8f1d7e2d24cefe5" upstream="master"/>
> <   <project name="openembedded/openembedded-core" path="layers/openembedded-core" remote="github" revision="8ab5b439ea82ac775494a0ce7a6f3615b61c94be" upstream="master"/>
> ---
> >   <project name="openembedded/meta-openembedded" path="layers/meta-openembedded" remote="github" revision="bb57bac845f3cd1634862fa9868bc8e294ba74a9" upstream="master"/>
> >   <project name="openembedded/openembedded-core" path="layers/openembedded-core" remote="github" revision="23f15c63777020f5d43b070a1eb2bcf246c19ff8" upstream="master"/>
>  
>  
> Other triage information: 
>  
> triage:
> ◾This build succeeded: http://xxxxx/job/xxx-master/559/
> ◾This build failed: http://xxxxx//job/xxx-master/560/
> ◾The only projects that added changes between these to versions are openembedded-core and meta-openembedded (not meta-linaro containing meta-optee which holds the breaking optee-client.bb recipe).
> ◾Pinning oe-core in default.xml manifest to build 559 pin builds successfully (leaving meta-oe at head of master).
> ◾Pinning meta-oe in default manifest to build 559 pin doesn't build successfully (leaving oe-core at head of master).
> ◾Numerous oe-core changes have been added wrt gcc compiler recipes. Likely that a compiler option has been added to treat certain warnings as errors causing the latest break.
> ◾Should notify linaro project of the break and see if they fix it, or whether oe-core changes accommodate (less likely).
>

jforissier · 2018-06-18T12:45:48Z

Hi,

I don't have GCC 8.1.1 so I can't reproduce the issue (I'm running Ubuntu 16.04.4). I suspect increasing the size of char prefixed[MAX_PRINT_SIZE] by 11 could very well fix the problem.

afaerber · 2018-06-18T16:19:01Z

Why 11? 255-246=9, which appears to assume that all preceding %s are empty and %d is only one char, otherwise it's at least 13 by my count. I rather feared that we may need to use %*s and pass a suitable length < 256-x there?

If needed, a Tumbleweed rootfs for chroot can be found here: https://download.opensuse.org/ports/aarch64/tumbleweed/images/openSUSE-Tumbleweed-ARM-JeOS.aarch64-rootfs.aarch64-Current.tar.xz (it may need a zypper up to update to gcc8)

jforissier · 2018-06-18T16:31:24Z

I should have said 10 = 266 - 256. Any chance you could try the following patch?

diff --git a/libteec/src/teec_trace.c b/libteec/src/teec_trace.c
index 78b79d6..54a53ec 100644
--- a/libteec/src/teec_trace.c
+++ b/libteec/src/teec_trace.c
@@ -73,7 +73,7 @@ int _dprintf(const char *function, int flen, int line, int level,
             const char *prefix, const char *fmt, ...)
 {
        char raw[MAX_PRINT_SIZE];
-       char prefixed[MAX_PRINT_SIZE];
+       char prefixed[MAX_PRINT_SIZE + 10];
        char *to_print = NULL;
        const char *func;
        int err;

scarabeusiv · 2018-06-18T19:35:29Z

That will not fix this error. For our openSUSE purposes the problem is default for Werror, which I just disabled -> solved as warning only for us.

simonqhughes · 2018-06-19T12:53:55Z

@scarabeusiv do you have a patch that we could use? We could use it temporarily before (@jforissier) a better fix is tested.

afaerber · 2018-06-19T13:04:01Z

No, it's a sed script in the spec file: https://build.opensuse.org/package/view_file/hardware:boot/optee-client/optee-client.spec?expand=1

omnium21 · 2018-06-21T15:33:44Z

@jforissier, I tried the patch and it didn't work, because the old length was still being passed into snprintf. so this patch worked:

diff --git a/libteec/src/teec_trace.c b/libteec/src/teec_trace.c
index 78b79d6..7901deb 100644
--- a/libteec/src/teec_trace.c
+++ b/libteec/src/teec_trace.c
@@ -73,7 +73,7 @@ int _dprintf(const char *function, int flen, int line, int level,
 	     const char *prefix, const char *fmt, ...)
 {
 	char raw[MAX_PRINT_SIZE];
-	char prefixed[MAX_PRINT_SIZE];
+	char prefixed[MAX_PRINT_SIZE + 10];
 	char *to_print = NULL;
 	const char *func;
 	int err;
@@ -106,7 +106,7 @@ int _dprintf(const char *function, int flen, int line, int level,
 		 */
 		int thread_id = syscall(SYS_gettid);	/* perf issue ? */
 
-		snprintf(prefixed, MAX_PRINT_SIZE,
+		snprintf(prefixed, MAX_PRINT_SIZE + 10,
 			 "%s [%d] %s:%s:%d: %s",
 			 trace_level_strings[level], thread_id, prefix, func,
 			 line, raw);

Of course, it's not nice to add 10 to two places, and saying as the define is MAX_PRINT_SIZE, we probably shouldn't be adding to it.

This patch looked simpler to me:

diff --git a/libteec/src/teec_trace.c b/libteec/src/teec_trace.c
index 78b79d6..4056661 100644
--- a/libteec/src/teec_trace.c
+++ b/libteec/src/teec_trace.c
@@ -72,7 +72,7 @@ static const char * const trace_level_strings[] = {
 int _dprintf(const char *function, int flen, int line, int level,
 	     const char *prefix, const char *fmt, ...)
 {
-	char raw[MAX_PRINT_SIZE];
+	char raw[MAX_PRINT_SIZE - 10];
 	char prefixed[MAX_PRINT_SIZE];
 	char *to_print = NULL;
 	const char *func;

However, 10 probably isn't the correct value anyway.

The prefix string could potentially be longer than 10 characters. It's likely to be 13 or more characters, based on this format string "%s [%d] %s:%s:%d: %s" and only if each variable is 1 character long in the final string.

But trace_level_strings can be 0 or 3 chars long; thread id is likely to be many characters long, potentially up to 11 character for a 32-bit int or 20 characters for a 64 bit int; prefix and func are be variable length strings passed in, so it might be better to use %{value}s to limit their length and take that into account; then there's the final line int, which again could be up to 20 characters long.

That makes 3 + 20 + prefix string + func string + 20 + 3 spaces + 3 colons + 2 brackets == 51 + prefix string + func string.

So how about this instead?

diff --git a/libteec/src/teec_trace.c b/libteec/src/teec_trace.c
index 78b79d6..440b579 100644
--- a/libteec/src/teec_trace.c
+++ b/libteec/src/teec_trace.c
@@ -72,7 +72,7 @@ static const char * const trace_level_strings[] = {
 int _dprintf(const char *function, int flen, int line, int level,
 	     const char *prefix, const char *fmt, ...)
 {
-	char raw[MAX_PRINT_SIZE];
+	char raw[MAX_PRINT_SIZE - 100];
 	char prefixed[MAX_PRINT_SIZE];
 	char *to_print = NULL;
 	const char *func;
@@ -107,7 +107,7 @@ int _dprintf(const char *function, int flen, int line, int level,
 		int thread_id = syscall(SYS_gettid);	/* perf issue ? */
 
 		snprintf(prefixed, MAX_PRINT_SIZE,
-			 "%s [%d] %s:%s:%d: %s",
+			 "%3s [%d] %24s:%24s:%d: %s",
 			 trace_level_strings[level], thread_id, prefix, func,
 			 line, raw);
 		to_print = prefixed;

I'm happy to submit a patch if you tell me how you'd prefer to do it?

simonqhughes · 2018-06-22T13:22:59Z

Please find attached a fix for this problem. I can submit as PR if given an indication its acceptable.

Its been through the following testing:

built with multiple other projects of multiple projects including optee_client and openembedded-core both at tip of master.

From 30dd2986fb64aba7ee78d4e231c344e2c39d7999 Mon Sep 17 00:00:00 2001
From: Simon Hughes <simon.hughes@arm.com>
Date: Thu, 21 Jun 2018 17:22:23 +0100
Subject: [PATCH] Fix for teec_trace.c snprintf -Werror=format-truncation=
 error.

Signed-off-by: Simon Hughes <simon.hughes@arm.com>
---
 libteec/src/teec_trace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/libteec/src/teec_trace.c b/libteec/src/teec_trace.c
index 78b79d6..c91bc43 100644
--- a/libteec/src/teec_trace.c
+++ b/libteec/src/teec_trace.c
@@ -106,7 +106,8 @@ int _dprintf(const char *function, int flen, int line, int level,
 		 */
 		int thread_id = syscall(SYS_gettid);	/* perf issue ? */
 
-		snprintf(prefixed, MAX_PRINT_SIZE,
+		int len = 0;
+		len  = snprintf(prefixed+len, MAX_PRINT_SIZE,
 			 "%s [%d] %s:%s:%d: %s",
 			 trace_level_strings[level], thread_id, prefix, func,
 			 line, raw);
-- 
2.7.4

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

GCC 8 format-truncation error #126

GCC 8 format-truncation error #126

afaerber commented Jun 17, 2018

afaerber commented Jun 17, 2018

simonqhughes commented Jun 18, 2018 •

edited

jforissier commented Jun 18, 2018

afaerber commented Jun 18, 2018

jforissier commented Jun 18, 2018

scarabeusiv commented Jun 18, 2018

simonqhughes commented Jun 19, 2018

afaerber commented Jun 19, 2018

omnium21 commented Jun 21, 2018

simonqhughes commented Jun 22, 2018 •

edited

omnium21 commented Jun 25, 2018

jforissier commented Jul 5, 2018

github-actions bot commented Dec 24, 2021

GCC 8 format-truncation error #126

GCC 8 format-truncation error #126

Comments

afaerber commented Jun 17, 2018

afaerber commented Jun 17, 2018

simonqhughes commented Jun 18, 2018 • edited

jforissier commented Jun 18, 2018

afaerber commented Jun 18, 2018

jforissier commented Jun 18, 2018

scarabeusiv commented Jun 18, 2018

simonqhughes commented Jun 19, 2018

afaerber commented Jun 19, 2018

omnium21 commented Jun 21, 2018

simonqhughes commented Jun 22, 2018 • edited

omnium21 commented Jun 25, 2018

jforissier commented Jul 5, 2018

github-actions bot commented Dec 24, 2021

simonqhughes commented Jun 18, 2018 •

edited

simonqhughes commented Jun 22, 2018 •

edited