linux-kernel - [PATCH perf/core v9 01/16] perf-symbol: Introduce filename_

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20160528151513.16098.97576.stgit@devbox>
Date:	Sun, 29 May 2016 00:15:13 +0900
From:	Masami Hiramatsu <mhiramat@...nel.org>
To:	Arnaldo Carvalho de Melo <acme@...nel.org>
Cc:	Masami Hiramatsu <mhiramat@...nel.org>,
	linux-kernel@...r.kernel.org, Namhyung Kim <namhyung@...nel.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...hat.com>,
	Hemant Kumar <hemant@...ux.vnet.ibm.com>,
	Ananth N Mavinakayanahalli <ananth@...ux.vnet.ibm.com>,
	Brendan Gregg <brendan.d.gregg@...il.com>
Subject: [PATCH perf/core v9 01/16] perf-symbol: Introduce filename__readable to check readability

Introduce filename__readable to check readability by opening
the file directly. Since the access(R_OK) just checks the
readability based on real UID/GID, it is ignored that the
effective UID/GID and capabilities for some special file (e.g.
/proc/kcore).
filename__readable() directly opens given file with O_RDONLY
so that the kernel checks it by effective UID/GID and capabilities.
---
 tools/perf/util/symbol.c |   32 ++++++++++++++++++++++----------
 1 file changed, 22 insertions(+), 10 deletions(-)

diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 54c4ff2..a469346 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1641,6 +1641,20 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
 	return ret;
 }
 
+/*
+ * Use open(O_RDONLY) to check readability directly instead of access(R_OK)
+ * since access(R_OK) only checks with real UID/GID but open() use effective
+ * UID/GID and actual capabilities (e.g. /proc/kcore requires CAP_SYS_RAWIO).
+ */
+static bool filename__readable(const char *file)
+{
+	int fd = open(file, O_RDONLY);
+	if (fd < 0)
+		return false;
+	close(fd);
+	return true;
+}
+
 static char *dso__find_kallsyms(struct dso *dso, struct map *map)
 {
 	u8 host_build_id[BUILD_ID_SIZE];
@@ -1668,7 +1682,6 @@ static char *dso__find_kallsyms(struct dso *dso, struct map *map)
 	/* Use /proc/kallsyms if possible */
 	if (is_host) {
 		DIR *d;
-		int fd;
 
 		/* If no cached kcore go with /proc/kallsyms */
 		d = opendir(path);
@@ -1677,16 +1690,15 @@ static char *dso__find_kallsyms(struct dso *dso, struct map *map)
 		closedir(d);
 
 		/*
-		 * Do not check the build-id cache, until we know we cannot use
-		 * /proc/kcore.
+		 * Do not check the build-id cache, unless we know we cannot use
+		 * /proc/kcore or module maps don't match to /proc/kallsyms.
+		 * To check readability of /proc/kcore, do not use access(R_OK)
+		 * since /proc/kcore requires CAP_SYS_RAWIO to read and access
+		 * can't check it.
 		 */
-		fd = open("/proc/kcore", O_RDONLY);
-		if (fd != -1) {
-			close(fd);
-			/* If module maps match go with /proc/kallsyms */
-			if (!validate_kcore_addresses("/proc/kallsyms", map))
-				goto proc_kallsyms;
-		}
+		if (filename__readable("/proc/kcore") &&
+		    !validate_kcore_addresses("/proc/kallsyms", map))
+			goto proc_kallsyms;
 
 		/* Find kallsyms in build-id cache with kcore */
 		if (!find_matching_kcore(map, path, sizeof(path)))