mknod02.c: Simplify and convert to new LTP API
PATCH v2 1/1 mknod02.c: Simplify and convert to new LTP API
Simply test when parent directory does not have set-group-ID bit set, new node gets GID from effective GID of the process and does not inherit the group ownership from its parent directory.
--- testcases/kernel/syscalls/mknod/mknod02.c | 316 +++------------------- 1 file changed, 36 insertions(+), 280 deletions(-)
testcases/kernel/syscalls/mknod/mknod02.c +36 -280
@@ -1,301 +1,57 @@ +// SPDX-License-Identifier: GPL-2.0-or-later /* -* * Copyright (c) International Business Machines Corp., 2001 -* -* This program is free software; you can redistribute it and/or modify -* it under the terms of the GNU General Public License as published by -* the Free Software Foundation; either version 2 of the License, or -* (at your option) any later version. -* -* This program is distributed in the hope that it will be useful, -* but WITHOUT ANY WARRANTY; without even the implied warranty of -* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See -* the GNU General Public License for more details. -* -* You should have received a copy of the GNU General Public License -* along with this program; if not, write to the Free Software -* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +* 07/2001 Ported by Wayne Boyer +* Copyright (c) 2023 SUSE LLC Avinesh Kumar <[email protected]> */ -/* -* Test Name: mknod02 -* -* Test Description: -* Verify that mknod(2) succeeds when used to create a filesystem -* node with set group-ID bit set on a directory without set group-ID bit set. -* The node created should have set group-ID bit set and its gid should be -* equal to that of its parent directory. -* -* Expected Result: -* mknod() should return value 0 on success and node created should have -* set group-ID bit set, its gid should be equal to that of its parent -* directory. -* -* Algorithm: -* Setup: -* Setup signal handling. -* Create temporary directory. -* Pause for SIGUSR1 if option specified. -* -* Test: -* Loop if the proper options are given. -* Execute system call -* Check return code, if system call failed (return=-1) -* Log the errno and Issue a FAIL message. -* Otherwise, -* Verify the Functionality of system call -* if successful, -* Issue Functionality-Pass message. -* Otherwise, -* Issue Functionality-Fail message. -* Cleanup: -* Print errno log and/or timing stats if options given -* Delete the temporary directory created. -* -* Usage: <for command-line> -* mknod02 [-c n] [-f] [-i n] [-I x] [-P x] [-t] -* where, -c n : Run n copies concurrently. -* -f : Turn off functionality Testing. -* -i n : Execute test n times. -* -I x : Execute test for x seconds. -* -P x : Pause for x seconds between iterations. -* -t : Turn on syscall timing. -* -* HISTORY -* 07/2001 Ported by Wayne Boyer -* -* RESTRICTIONS: -* This test should be run by 'super-user' (root) only. +/*\ +* [Description] * +* Verify that if mknod(2) creates a filesystem node in a directory which +* does not have the set-group-ID bit set, new node will not inherit the +* group ownership from its parent directory and its group ID will be the +* effective group ID of the process. */ -#include <stdio.h> -#include <stdlib.h> -#include <unistd.h> -#include <errno.h> -#include <string.h> -#include <signal.h> #include <pwd.h> -#include <sys/types.h> -#include <sys/stat.h> - -#include "test.h" -#include "safe_macros.h" +#include "tst_test.h" -#define LTPUSER "nobody" -#define MODE_RWX S_IFIFO | S_IRWXU | S_IRWXG | S_IRWXO -#define MODE_SGID S_IFIFO | S_ISGID | S_IRWXU | S_IRWXG | S_IRWXO -#define DIR_TEMP "testdir_2" -#define TNODE "tnode_%d" +#define MODE_DIR 0777 +#define MODE1 0010777 +#define MODE_SGID 02000 -struct stat buf;/* struct. to hold stat(2) o/p contents */ -struct passwd *user1;/* struct. to hold getpwnam(3) o/p contents */ +#define TEMP_DIR "testdir" +#define TEMP_NODE "testnode" -char *TCID ="mknod02" ; -int TST_TOTAL =1 ; -char node_name[PATH_MAX ];/* buffer to hold node name created */ +static struct stat buf; +static struct passwd *user_nobody; +static gid_t gid_nobody; -gid_t group1_gid, group2_gid, mygid;/* user and process group id's */ -uid_t save_myuid, user1_uid;/* user and process user id's */ -pid_t mypid;/* process id */ - -void setup ();/* setup function for the test */ -void cleanup ();/* cleanup function for the test */ - -int main (int ac,char **av) +static void setup (void ) { -int lc; -int fflag; - -tst_parse_opts (ac, av,NULL ,NULL ); - -setup (); - -for (lc =0 ;TEST_LOOPING (lc); lc++) { - - tst_count =0 ; - -/* -* Attempt to create a filesystem node with group-id bit set -* on a directory without group id bit set such that, -* the node created by mknod(2) should have group-id (sgid) -* bit set and node's gid should be equal to that of its -* parent directory. -*/ -TEST (mknod (node_name,MODE_SGID ,0 )); - -/* Check return code from mknod(2) */ -if (TEST_RETURN ==-1 ) { -tst_resm (TFAIL , -"mknod(%s, %#o, 0) failed, errno=%d : %s" , - node_name,MODE_SGID ,TEST_ERRNO , -strerror (TEST_ERRNO )); -continue ; - } -/* Set the functionality flag */ - fflag =1 ; - -/* Check for node's creation */ -if (stat (node_name, &buf) <0 ) { -tst_resm (TFAIL ,"stat() of %s failed, errno:%d" , - node_name,TEST_ERRNO ); - -/* unset functionality flag */ - fflag =0 ; - } - -/* Verify mode permissions of node */ -if (!(buf.st_mode &S_ISGID )) { -tst_resm (TFAIL ,"%s: Incorrect modes, setgid " -"bit not set" , node_name); -/* unset flag as functionality fails */ - fflag =0 ; - } - -/* Verify group ID of node */ -if (buf.st_gid != mygid) { -tst_resm (TFAIL ,"%s: Incorrect group" , - node_name); -/* unset flag as functionality fails */ - fflag =0 ; - } -if (fflag) { -tst_resm (TPASS ,"Functionality of mknod(%s, " -"%#o, 0) successful" , - node_name,MODE_SGID ); - } - -/* Remove the node for the next go `round */ -if (unlink (node_name) ==-1 ) { -tst_resm (TWARN ,"unlink(%s) failed, errno:%d %s" , - node_name, errno,strerror (errno)); - } - } + user_nobody =SAFE_GETPWNAM ("nobody" ); + gid_nobody = user_nobody->pw_gid; -/* Change the directory back to temporary directory */ -SAFE_CHDIR (cleanup,".." ); - -/* -* Invoke cleanup() to delete the test directories created -* in the setup() and exit main(). -*/ -cleanup (); - -tst_exit (); +SAFE_MKDIR (TEMP_DIR ,MODE_DIR ); +SAFE_CHOWN (TEMP_DIR ,-1 , gid_nobody); } -/* -* setup(void) - performs all ONE TIME setup for this test. -* Exit the test program on receipt of unexpected signals. -* Create a temporary directory used to hold test directories created -* and change the directory to it. -* Verify that pid of process executing the test is root. -* Create a test directory on temporary directory and set the ownership -* of test directory to ltp user and process. -* Set the effective uid/gid of the process to that of ltp user. -*/ -void setup (void ) +static void run (void ) { -tst_require_root (); - -/* Capture unexpected signals */ -tst_sig (NOFORK ,DEF_HANDLER , cleanup); - -TEST_PAUSE ; - -/* Make a temp dir and cd to it */ -tst_tmpdir (); - -/* fix permissions on the tmpdir */ -if (chmod ("." ,0711 ) !=0 ) { -tst_brkm (TBROK , cleanup,"chmod() failed" ); - } - -/* Save the real user id of the current test process */ - save_myuid =getuid (); - -/* Save the process id of the current test process */ - mypid =getpid (); - -/* Get the node name to be created in the test */ -sprintf (node_name,TNODE , mypid); - -/* Get the uid/gid of ltpuser */ -if ((user1 =getpwnam (LTPUSER )) ==NULL ) { -tst_brkm (TBROK |TERRNO , cleanup, -"Couldn't determine if %s was in /etc/passwd" , -LTPUSER ); - } - user1_uid = user1->pw_uid; - group1_gid = user1->pw_gid; - -/* Get the effective group id of the test process */ - group2_gid =getegid (); +SAFE_CHDIR (TEMP_DIR ); +TST_EXP_PASS (mknod (TEMP_NODE ,MODE1 ,0 ),"mknod(%s, %o, 0)" ,TEMP_NODE ,MODE1 ); -/* -* Create a test directory under temporary directory with the -* specified mode permissions, with uid/gid set to that of guest -* user and the test process. -*/ -SAFE_MKDIR (cleanup,DIR_TEMP ,MODE_RWX ); -SAFE_CHOWN (cleanup,DIR_TEMP , user1_uid, group2_gid); +SAFE_STAT (TEMP_NODE , &buf); +TST_EXP_EQ_LI (buf.st_gid,0 ); -/* -* Verify that test directory created with expected permission modes -* and ownerships. -*/ -SAFE_STAT (cleanup,DIR_TEMP , &buf); - -/* Verify modes of test directory */ -if (buf.st_mode &S_ISGID ) { -tst_brkm (TBROK , cleanup, -"%s: Incorrect modes, setgid bit set" ,DIR_TEMP ); - } - -/* Verify group ID of test directory */ -if (buf.st_gid != group2_gid) { -tst_brkm (TBROK , cleanup,"%s: Incorrect group" ,DIR_TEMP ); - } - -/* -* Set the effective group id and user id of the test process -* to that of guest user. -*/ -SAFE_SETGID (cleanup, group1_gid); -if (setreuid (-1 , user1_uid) <0 ) { -tst_brkm (TBROK , cleanup, -"Unable to set process uid to that of ltp user" ); - } - -/* Save the real group ID of the current process */ - mygid =getgid (); - -/* Change directory to DIR_TEMP */ -SAFE_CHDIR (cleanup,DIR_TEMP ); +SAFE_UNLINK (TEMP_NODE ); +SAFE_CHDIR (".." ); } -/* -* cleanup() - Performs all ONE TIME cleanup for this test at -* completion or premature exit. -* Print test timing stats and errno log if test executed with options. -* Restore the real/effective user id of the process changed during -* setup(). -* Remove temporary directory and sub-directories/files under it -* created during setup(). -* Exit the test program with normal exit code. -*/ -void cleanup (void ) -{ - -/* -* Restore the effective uid of the process changed in the -* setup(). -*/ -if (setreuid (-1 , save_myuid) <0 ) { -tst_brkm (TBROK ,NULL , -"resetting process real/effective uid failed" ); - } - -tst_rmdir (); - -} +static struct tst_test test = { + .setup = setup, + .test_all = run, + .needs_root =1 , + .needs_tmpdir =1 +};
Signature · 5 lines
2.39.2 -- Mailing list info: https://lists.linux.it/listinfo/ltp
Hi Avinesh,
Quoted 3 lines
Simply test when parent directory does not have set-group-ID bit set, new node gets GID from effective GID of the process and does not inherit the group ownership from its parent directory.
Few comments below.
Quoted 2 lines
testcases/kernel/syscalls/mknod/mknod02.c | 316 +++------------------- 1 file changed, 36 insertions(+), 280 deletions(-)
...
Quoted 7 lines
+/*\
+ * [Description]
*
+ * Verify that if mknod(2) creates a filesystem node in a directory which
+ * does not have the set-group-ID bit set, new node will not inherit the
+ * group ownership from its parent directory and its group ID will be the
+ * effective group ID of the process.
@Cyril I wonder if it'd be good to test this on all_filesystems. Are we trying to use use all_filesystems = 1 when subject of testing is using VFS or the opposite? (kernel docs mentions "VFS system calls open(2), stat(2), read(2), write(2), chmod(2)". It also mentions locking [2]).
BTW looking what has mknod in vfs, it's just nfsd and 9p (none of them are used in all_filesystems):
$ git grep mknod $(git ls-files fs/|grep -i vfs) fs/9p/vfs_inode.c: * for mknod(2). fs/9p/vfs_inode.c: * v9fs_vfs_mknod - create a special file fs/9p/vfs_inode.c:v9fs_vfs_mknod(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: return v9fs_vfs_mknod_dotl(idmap, dir, dentry, omode, 0); fs/9p/vfs_inode_dotl.c: * v9fs_vfs_mknod_dotl - create a special file fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: p9_debug(P9_DEBUG_VFS, "Failed to get acl values in mknod %d\n", fs/9p/vfs_inode_dotl.c: err = p9_client_mknod_dotl(dfid, name, mode, rdev, gid, &qid); fs/9p/vfs_inode_dotl.c: .mknod = v9fs_vfs_mknod_dotl, fs/nfsd/vfs.c: host_err = vfs_mknod(&nop_mnt_idmap, dirp, dchild,
Quoted 1 line
*/
...
Quoted 2 lines
+static void run(void)
{
...
testcases/kernel/syscalls/mknod/mknod02.c line 43
Quoted 2 lines
testcases/kernel/syscalls/mknod/mknod02.c +2
-0+SAFE_CHDIR (TEMP_DIR ); +TST_EXP_PASS (mknod (TEMP_NODE ,MODE1 ,0 ),"mknod(%s, %o, 0)" ,TEMP_NODE ,MODE1 );
IMHO this simple form will print the same info, I suggest to use it.
TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0));
testcases/kernel/syscalls/mknod/mknod02.c line 46
Quoted 9 lines
testcases/kernel/syscalls/mknod/mknod02.c +2
-7-/* -* Create a test directory under temporary directory with the -* specified mode permissions, with uid/gid set to that of guest -* user and the test process. -*/ -SAFE_MKDIR (cleanup,DIR_TEMP ,MODE_RWX ); -SAFE_CHOWN (cleanup,DIR_TEMP , user1_uid, group2_gid); +SAFE_STAT (TEMP_NODE , &buf); +TST_EXP_EQ_LI (buf.st_gid,0 );
...
Diff for using all_filesystems (I'm not sure myself).
Kind regards,
Petr
[1] https://www.kernel.org/doc/html/next/filesystems/vfs.html
[2] https://www.kernel.org/doc/html/next/filesystems/locking.html
+++ testcases/kernel/syscalls/mknod/mknod02.c
@@ -24,6 +24,8 @@
#define TEMP_DIR "testdir"
#define TEMP_NODE "testnode"
+#define MNTPOINT "mntpoint"
+
static struct stat buf;
static struct passwd *user_nobody;
static gid_t gid_nobody;
@@ -40,18 +42,28 @@ static void setup(void)
static void run(void)
{
SAFE_CHDIR(TEMP_DIR);
- TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0), "mknod(%s, %o, 0)", TEMP_NODE, MODE1);
+ TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0));
SAFE_STAT(TEMP_NODE, &buf);
TST_EXP_EQ_LI(buf.st_gid, 0);
SAFE_UNLINK(TEMP_NODE);
SAFE_CHDIR("..");
+
+}
+
+static void cleanup(void)
+{
+ SAFE_RMDIR(TEMP_DIR);
}
static struct tst_test test = {
.setup = setup,
+ .cleanup = cleanup,
.test_all = run,
.needs_root = 1,
- .needs_tmpdir = 1
+ .needs_tmpdir = 1,
+ .mount_device = 1,
+ .mntpoint = MNTPOINT,
+ .all_filesystems = 1,
};
Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
Hi Petr,
Quoted 53 lines · IST Petr Vorel
Hi Avinesh,
Quoted 3 lines
Simply test when parent directory does not have set-group-ID bit set, new node gets GID from effective GID of the process and does not inherit the group ownership from its parent directory.
Reviewed-by Petr Vorel <[email protected]>Few comments below.
Quoted 2 lines
testcases/kernel/syscalls/mknod/mknod02.c | 316 +++------------------- 1 file changed, 36 insertions(+), 280 deletions(-)...
Quoted 9 lines
+/*\
+ * [Description]*
+ * Verify that if mknod(2) creates a filesystem node in a directory which
+ * does not have the set-group-ID bit set, new node will not inherit the
+ * group ownership from its parent directory and its group ID will be the
+ * effective group ID of the process.@Cyril I wonder if it'd be good to test this on
all_filesystems. Are we trying to use useall_filesystems= 1 when subject of testing is using VFS or the opposite? (kernel docs mentions "VFS system calls open(2), stat(2), read(2), write(2), chmod(2)". It also mentions locking [2]).BTW looking what has mknod in vfs, it's just nfsd and 9p (none of them are used in
all_filesystems):$ git grep mknod $(git ls-files fs/|grep -i vfs) fs/9p/vfs_inode.c: * for mknod(2). fs/9p/vfs_inode.c: * v9fs_vfs_mknod - create a special file fs/9p/vfs_inode.c:v9fs_vfs_mknod(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: return v9fs_vfs_mknod_dotl(idmap, dir, dentry, omode, 0); fs/9p/vfs_inode_dotl.c: * v9fs_vfs_mknod_dotl - create a special file fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: p9_debug(P9_DEBUG_VFS, "Failed to get acl values in mknod %d\n", fs/9p/vfs_inode_dotl.c: err = p9_client_mknod_dotl(dfid, name, mode, rdev, gid, &qid); fs/9p/vfs_inode_dotl.c: .mknod = v9fs_vfs_mknod_dotl, fs/nfsd/vfs.c: host_err = vfs_mknod(&nop_mnt_idmap, dirp, dchild,Quoted 1 line
*/
...
Quoted 8 lines
Quoted 3 lines
+static void run(void)
{
...
Review
Quoted 2 lines
(quoted diff) +2
-0+ SAFE_CHDIR(TEMP_DIR); + TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0), "mknod(%s, %o, 0)",
TEMP_NODE,
Quoted 16 lines
Quoted 1 line
MODE1);
IMHO this simple form will print the same info, I suggest to use it.
TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0));Quoted 9 lines
testcases/kernel/syscalls/mknod/mknod02.c +2
-7-/* -* Create a test directory under temporary directory with the -* specified mode permissions, with uid/gid set to that of guest -* user and the test process. -*/ -SAFE_MKDIR (cleanup,DIR_TEMP ,MODE_RWX ); -SAFE_CHOWN (cleanup,DIR_TEMP , user1_uid, group2_gid); +SAFE_STAT (TEMP_NODE , &buf); +TST_EXP_EQ_LI (buf.st_gid,0 );...
Quoted 4 lines
Diff for using
all_filesystems(I'm not sure myself).Kind regards,
Petr
Thanks for reviewing, I agree with your proposed changes if we are going to enable if for all filesystems, test executes fine with all_filesystems=1 in my setup.
Quoted 15 lines
[1] https://www.kernel.org/doc/html/next/filesystems/vfs.html
[2] https://www.kernel.org/doc/html/next/filesystems/locking.html+++
testcases/kernel/syscalls/mknod/mknod02.c(quoted diff) +2
-0@@ -24,6 +24,8 @@ #define TEMP_DIR "testdir" #define TEMP_NODE "testnode" +#define MNTPOINT "mntpoint" + static struct stat buf; static struct passwd *user_nobody; static gid_t gid_nobody;
Review static void setup(void) · line 43
Quoted 4 lines
(quoted diff) +0
-1@@ -40,18 +42,28 @@ static void setup(void) static void run(void) { SAFE_CHDIR(TEMP_DIR); - TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0), "mknod(%s, %o, 0)",
TEMP_NODE,
Quoted 16 lines
(quoted diff) +6
-0MODE1); + TST_EXP_PASS(mknod(TEMP_NODE, MODE1, 0)); SAFE_STAT(TEMP_NODE, &buf); TST_EXP_EQ_LI(buf.st_gid, 0); SAFE_UNLINK(TEMP_NODE); SAFE_CHDIR(".."); + +} + +static void cleanup(void) +{ + SAFE_RMDIR(TEMP_DIR); } static struct tst_test test = {
Review static void setup(void)
Quoted 10 lines
(quoted diff) +5
-1.setup = setup, + .cleanup = cleanup, .test_all = run, .needs_root = 1, - .needs_tmpdir = 1 + .needs_tmpdir = 1, + .mount_device = 1, + .mntpoint = MNTPOINT, + .all_filesystems = 1, };
Regards,
Avinesh
Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
Hello,
Quoted 30 lines · Petr Vorel
Hi Avinesh,
Quoted 3 lines
Simply test when parent directory does not have set-group-ID bit set, new node gets GID from effective GID of the process and does not inherit the group ownership from its parent directory.
Reviewed-by Petr Vorel <[email protected]>Few comments below.
Quoted 2 lines
testcases/kernel/syscalls/mknod/mknod02.c | 316 +++------------------- 1 file changed, 36 insertions(+), 280 deletions(-)...
Quoted 7 lines
+/*\
+ * [Description]
*
+ * Verify that if mknod(2) creates a filesystem node in a directory which
+ * does not have the set-group-ID bit set, new node will not inherit the
+ * group ownership from its parent directory and its group ID will be the
+ * effective group ID of the process.@Cyril I wonder if it'd be good to test this on
all_filesystems. Are we trying to use useall_filesystems= 1 when subject of testing is using VFS or the opposite? (kernel docs mentions "VFS system calls open(2), stat(2), read(2), write(2), chmod(2)". It also mentions locking [2]).BTW looking what has mknod in vfs, it's just nfsd and 9p (none of them are used in
all_filesystems):
Quoted 15 lines
$ git grep mknod $(git ls-files fs/|grep -i vfs) fs/9p/vfs_inode.c: * for mknod(2). fs/9p/vfs_inode.c: * v9fs_vfs_mknod - create a special file fs/9p/vfs_inode.c:v9fs_vfs_mknod(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode.c: .mknod = v9fs_vfs_mknod, fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: return v9fs_vfs_mknod_dotl(idmap, dir, dentry, omode, 0); fs/9p/vfs_inode_dotl.c: * v9fs_vfs_mknod_dotl - create a special file fs/9p/vfs_inode_dotl.c:v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, fs/9p/vfs_inode_dotl.c: p9_debug(P9_DEBUG_VFS, "Failed to get acl values in mknod %d\n", fs/9p/vfs_inode_dotl.c: err = p9_client_mknod_dotl(dfid, name, mode, rdev, gid, &qid); fs/9p/vfs_inode_dotl.c: .mknod = v9fs_vfs_mknod_dotl, fs/nfsd/vfs.c: host_err = vfs_mknod(&nop_mnt_idmap, dirp, dchild,
It is defined for all major file systems on the directory inode_operations struct (e.g. search btrfs_mknod, xfs_vn_mknod, ext4_mknod).
If you setup clangd and provide a reasonabl compile_commands.json (i.e. relatively large kernel config) you can search for places where struct inode_operations.mk_nod is set.
So I'll go ahead and try to merge it.
Signature · 5 lines
Thank you, Richard. -- Mailing list info: https://lists.linux.it/listinfo/ltp