lack of ENAMETOOLONG testcases for pathnames longer than PATH_MAX
Al Viro
maintainerPATCH 1/1 lack of ENAMETOOLONG testcases for pathnames longer than PATH_MAX
There are different causes of ENAMETOOLONG. It might come from
filesystem rejecting an excessively long pathname component, but there's also "pathname is longer than PATH_MAX bytes, including terminating NUL" and that doesn't get checked anywhere.
Ran into that when a braino in kernel patch broke that logics
(ending up with cutoff too low) and that didn't get caught by LTP run.
Patch below adds the checks to one of the tests that do deal
with the other source of ENAMETOOLONG; it almost certainly not the right use of infrastructure, though.
---
testcases/kernel/syscalls/chdir/chdir04.c +10 -1
@@ -11,6 +11,8 @@#include "tst_test.h" static char long_dir[] ="abcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyz" ; +static char long_path[PATH_MAX +1 ]; +static char shorter_path[PATH_MAX ];static char noexist_dir[] ="noexistdir" ; static struct tcase {
@@ -20,16 +22,23 @@ static struct tcase { {long_dir, ENAMETOOLONG}, {noexist_dir, ENOENT}, {0, EFAULT}, // bad_addr + {long_path, ENAMETOOLONG}, + {shorter_path, 0}, };static void verify_chdir (unsigned int i) { -TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +if (tcases[i].exp_errno) +TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +else +TST_EXP_PASS (chdir (tcases[i].dir),"chdir()" ); }static void setup (void ) { tcases[2 ].dir =tst_get_bad_addr (NULL ); +memset (long_path,'/' ,PATH_MAX ); +memset (shorter_path,'/' ,PATH_MAX -1 ); } static struct tst_test test = {
Hi!
Quoted 11 lines · Al Viro
There are different causes of ENAMETOOLONG. It might come from
filesystem rejecting an excessively long pathname component, but there's also "pathname is longer thanPATH_MAXbytes, including terminating NUL" and that doesn't get checked anywhere.Ran into that when a braino in kernel patch broke that logics
(ending up with cutoff too low) and that didn't get caught by LTP run.Patch below adds the checks to one of the tests that do deal
with the other source of ENAMETOOLONG; it almost certainly not the right use of infrastructure, though.
The description is not well formatted, spaces at the beginning of the phrases should be removed.
Also, we can make it slightly more clear, by saying that error can be caused by a path name that is bigger than NAME_MAX, if relative, or bigger than PATH_MAX, if absolute (when we use '/').
In this test we only verifies if relative paths are longer than NAME_MAX (we give 273 bytes instead of 255 max), but we don't test if path name is bigger than PATH_MAX.
We should correctly distinguish these two cases inside the test with proper names as well. Check below..
Quoted 8 lines
Signed-off-by Al Viro <[email protected]>---
testcases/kernel/syscalls/chdir/chdir04.c +0
-0
testcases/kernel/syscalls/chdir/chdir04.c line 16
Quoted 6 lines
testcases/kernel/syscalls/chdir/chdir04.c +2
-0@@ -11,6 +11,8 @@#include "tst_test.h" static char long_dir[] ="abcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyz" ; +static char long_path[PATH_MAX +1 ]; +static char shorter_path[PATH_MAX ];static char noexist_dir[] ="noexistdir" ;
When it comes to syscall testing, it's better to use guarded buffers. This is easy to do: please check tst_test.bufs usage in here:
Many old tests are not using these buffers, but it's better to introduce them when a test is refactored or fixed, like in this case.
You need to define:
static char *long_rel_path;
static char *long_abs_path;
...
static void setup(void) {
..
// initialize long_rel_path content
// initialize long_abs_path content
}static struct tst_test test = {
..
.bufs = (struct tst_buffer []) {
{&long_rel_path, .size = NAME_MAX + 10},
{&long_abs_path, .size = PATH_MAX + 10},
{}
}
};Quoted 7 lines
testcases/kernel/syscalls/chdir/chdir04.c +1
-0static struct tcase {@@ -20,16 +22,23 @@ static struct tcase { {long_dir, ENAMETOOLONG}, {noexist_dir, ENOENT}, {0, EFAULT}, // bad_addr + {long_path, ENAMETOOLONG},
testcases/kernel/syscalls/chdir/chdir04.c static struct tcase { · line 34
Quoted 10 lines
testcases/kernel/syscalls/chdir/chdir04.c +5
-1+ {shorter_path, 0}, };static void verify_chdir (unsigned int i) { -TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +if (tcases[i].exp_errno) +TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +else +TST_EXP_PASS (chdir (tcases[i].dir),"chdir()" );
In this test we only verify errors, so TST_EXP_PASS is not needed.
Signature · 3 lines
Andrea Cervesato SUSE QE Automation Engineer Linux [email protected]
I forgot to mention that guarded buffers should be introduced for noexist_dir as well.
Kind regards,
Signature · 3 lines
Andrea Cervesato SUSE QE Automation Engineer Linux [email protected]
Al Viro
maintainerQuoted 15 lines · Andrea Cervesato
Hi!
Quoted 11 lines · Al Viro
There are different causes of ENAMETOOLONG. It might come from
filesystem rejecting an excessively long pathname component, but there's also "pathname is longer thanPATH_MAXbytes, including terminating NUL" and that doesn't get checked anywhere.Ran into that when a braino in kernel patch broke that logics
(ending up with cutoff too low) and that didn't get caught by LTP run.Patch below adds the checks to one of the tests that do deal
with the other source of ENAMETOOLONG; it almost certainly not the right use of infrastructure, though.
Quoted 6 lines
The description is not well formatted, spaces at the beginning of the phrases should be removed.
Also, we can make it slightly more clear, by saying that error can be caused by a path name that is bigger than
NAME_MAX, if relative, or bigger thanPATH_MAX, if absolute (when we use '/').
Huh? Absolute pathname is the one that _starts_ with '/'; e.g. "../include" is relative, not absolute, despite having a '/' in it.
Quoted 34 lines
In this test we only verifies if relative paths are longer than
NAME_MAX(we give 273 bytes instead of 255 max), but we don't test if path name is bigger thanPATH_MAX.We should correctly distinguish these two cases inside the test with proper names as well. Check below..
Many old tests are not using these buffers, but it's better to introduce them when a test is refactored or fixed, like in this case.
You need to define:
static char
*long_rel_path;
static char*long_abs_path;...
static void setup(void) { .. // initialize long_rel_path content // initialize long_abs_path content }static struct tst_test test = { .. .bufs = (struct tst_buffer []) { {&long_rel_path, .size = NAME_MAX + 10}, {&long_abs_path, .size = PATH_MAX + 10}, {} } };
Quoted 7 lines
testcases/kernel/syscalls/chdir/chdir04.c static struct tcase { · line 34
Quoted 5 lines
testcases/kernel/syscalls/chdir/chdir04.c +4
-1-TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +if (tcases[i].exp_errno) +TST_EXP_FAIL (chdir (tcases[i].dir), tcases[i].exp_errno,"chdir()" ); +else +TST_EXP_PASS (chdir (tcases[i].dir),"chdir()" );In this test we only verify errors, so
TST_EXP_PASSis not needed.
Er... Intent was to verify two things: that anything longer than PATH_MAX triggers ENAMETOOLONG, but anything up to PATH_MAX does not. Having a pathname of exactly 4095 '/' (or interleaved . and / in the same amount, etc.) be rejected with ENAMETOOLONG is just as much of a failure as not triggering ENAMETOOLONG on anything longer...
FWIW, I considered something like
mkdir("subdirectory", 0700);concatenating enough copies of "subdirectory/../" to get just under PATH_MAX and appending "././././././././" to the end, so that truncation to PATH_MAX and to PATH_MAX-1 would both be otherwise valid paths; decided that it's better to keep it simpler - a pile of slashes is easier to produce and would resolve to a valid directory if not for the total length restrictions.
Quoted 5 lines
Er... Intent was to verify two things: that anything longer than
PATH_MAXtriggers ENAMETOOLONG, but anything up toPATH_MAXdoes not. Having a pathname of exactly 4095 '/' (or interleaved . and / in the same amount, etc.) be rejected with ENAMETOOLONG is just as much of a failure as not triggering ENAMETOOLONG on anything longer...
In this case we need a new test verifying that PATH_MAX is actually handled well, as Cyril suggested. But in this test we should only verifying errors.
Quoted 8 lines
FWIW, I considered something like
mkdir("subdirectory", 0700);concatenating enough copies of "subdirectory/../" to get just under
PATH_MAXand appending "././././././././" to the end, so that truncation toPATH_MAXand to PATH_MAX-1 would both be otherwise valid paths; decided that it's better to keep it simpler - a pile of slashes is easier to produce and would resolve to a valid directory if not for the total length restrictions.
It's up to you how you create the string that will trigger the error. Also, you probably need to take a look at tst_test.needs_tmpdir.
Signature · 3 lines
Andrea Cervesato SUSE QE Automation Engineer Linux [email protected]
Hi!
Quoted 4 lines
There are different causes of ENAMETOOLONG. It might come from
filesystem rejecting an excessively long pathname component, but there's also "pathname is longer thanPATH_MAXbytes, including terminating NUL" and that doesn't get checked anywhere.
We do have a couple of tests that checks that names over PATH_MAX are rejected, there is no reason to add these kind of tests, however I do not think that we tests that check that names that are just under the limit work fine, that needs to be added.
Quoted 16 lines
Signed-off-by Al Viro <[email protected]>---
testcases/kernel/syscalls/chdir/chdir04.c +2
-0@@ -11,6 +11,8 @@#include "tst_test.h" static char long_dir[] ="abcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyzabcdefghijklmnopqrstmnopqrstuvwxyz" ; +static char long_path[PATH_MAX +1 ]; +static char shorter_path[PATH_MAX ];static char noexist_dir[] ="noexistdir" ; static struct tcase {
testcases/kernel/syscalls/chdir/chdir04.c static struct tcase { · line 25
Quoted 4 lines
testcases/kernel/syscalls/chdir/chdir04.c +1
-0@@ -20,16 +22,23 @@ static struct tcase { {long_dir, ENAMETOOLONG}, {noexist_dir, ENOENT}, {0, EFAULT}, // bad_addr + {long_path, ENAMETOOLONG},
This test already exists in the form of long_dir just three lines above.
testcases/kernel/syscalls/chdir/chdir04.c static struct tcase { · line 26
Quoted 1 line
testcases/kernel/syscalls/chdir/chdir04.c +1
-0+ {shorter_path, 0},
What about we add a separate test (chdir02.c) for paths that shouldn't be rejected. Something as:
char path[PATH_MAX];
int i;
...
for (i = 1; i < PATH_MAX; i++) {
memset(path, 0, sizeof(path));
memset(path, '/', i);
TST_EXP_PASS(chdir(path), "chdir() len=%i", i);
}...
That would make sure that all lenghts of paths that are valid are accepted.
Signature · 2 lines
Cyril Hrubis [email protected]