Skip to contentSkip to conversation

crypto/af_alg02: fixed read() being stuck

PATCH Superseded Christian Amann 3 participants 4 messages May 7, 2019

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 crypto/af_alg02: fixed read() being stuck

On kernels < 4.14 (missing commit 2d97591ef43d) reading from the request socket does not return and the testcase does not finish.

This fix moves the logic to a child thread in order for the parent to handle the timeout and report a message to the user.

---
 testcases/kernel/crypto/Makefile   |  2 ++
 testcases/kernel/crypto/af_alg02.c | 36 +++++++++++++++++++++++++++++++++++-
 2 files changed, 37 insertions(+), 1 deletion(-)
testcases/kernel/crypto/Makefile +2 -0
index 76f9308c2..6547e1cb6 100644
@@ -20,3 +20,5 @@ include $(top_srcdir)/include/mk/testcases.mk
 CFLAGS			+= -D_GNU_SOURCE
 
 include $(top_srcdir)/include/mk/generic_leaf_target.mk
+
+af_alg02: CFLAGS += -pthread
testcases/kernel/crypto/af_alg02.c +35 -1
index a9e820423..056511993 100644
@@ -7,12 +7,23 @@
  * Regression test for commit ecaaab564978 ("crypto: salsa20 - fix
  * blkcipher_walk API usage"), or CVE-2017-17805.  This test verifies that an
  * empty message can be encrypted with Salsa20 without crashing the kernel.
+ *
+ * read() fix:
+ * Calls read() in child thread in order to manually kill it after timeout.
+ * With kernels missing commit 2d97591ef43d ("crypto: af_alg - consolidation
+ * of duplicate code") read() does not return.
  */
 
 #include "tst_test.h"
 #include "tst_af_alg.h"
+#include "tst_safe_pthread.h"
+#include <pthread.h>
+#include <time.h>
+#include <errno.h>
 
-static void run(void)
+#define VERIFY_TIMEOUT (time(NULL) + 5)
+
+void *verify_encrypt(void *arg)
 {
 	char buf[16];
 	int reqfd = tst_alg_setup_reqfd("skcipher", "salsa20", NULL, 16);
@@ -22,6 +33,29 @@ static void run(void)
 		tst_res(TPASS, "Successfully \"encrypted\" an empty message");
 	else
 		tst_res(TBROK, "read() didn't return 0");
+	return arg;
+}
+
+static void run(void)
+{
+	pthread_t thr;
+	int join_ret;
+	struct timespec read_timeout;
+
+	read_timeout.tv_sec  = VERIFY_TIMEOUT;
+	read_timeout.tv_nsec = 0;
+
+	SAFE_PTHREAD_CREATE(&thr, NULL, verify_encrypt, NULL);
+	join_ret = pthread_timedjoin_np(thr, NULL, &read_timeout);
+
+	if (join_ret != 0) {
+		if (join_ret == ETIMEDOUT)
+			tst_brk(TBROK,
+				"Timed out while reading from request socket.");
+		else
+			tst_brk(TBROK | TTERRNO,
+				"Error while joining child thread");
+	}
 }
 
 static struct tst_test test = {
Signature · 1 line
2.16.4
#1 lore raw

PATCH 1/1 crypto/af_alg02: fixed read() being stuck

Hi Christian,

Quoted 48 lines · Christian Amann

On kernels < 4.14 (missing commit 2d97591ef43d) reading from the request socket does not return and the testcase does not finish.

This fix moves the logic to a child thread in order for the parent to handle the timeout and report a message to the user.

---
 testcases/kernel/crypto/Makefile   |  2 ++
 testcases/kernel/crypto/af_alg02.c | 36
+++++++++++++++++++++++++++++++++++-
 2 files changed, 37 insertions(+), 1 deletion(-)
testcases/kernel/crypto/Makefile +2 -0
index 76f9308c2..6547e1cb6 100644
@@ -20,3 +20,5 @@ include $(top_srcdir)/include/mk/testcases.mk
 CFLAGS                 += -D_GNU_SOURCE
 
 include $(top_srcdir)/include/mk/generic_leaf_target.mk
+
+af_alg02: CFLAGS += -pthread
testcases/kernel/crypto/af_alg02.c +0 -0
index a9e820423..056511993 100644
@@ -7,12 +7,23 @@
  * Regression test for commit ecaaab564978 ("crypto: salsa20 - fix
  * blkcipher_walk API usage"), or CVE-2017-17805.  This test verifies

that an
* empty message can be encrypted with Salsa20 without crashing the
kernel.
+ *
+ * read() fix:
+ * Calls read() in child thread in order to manually kill it after timeout.
+ * With kernels missing commit 2d97591ef43d ("crypto: af_alg - consolidation
+ * of duplicate code") read() does not return.
*/

#include "tst_test.h"
#include "tst_af_alg.h"

testcases/kernel/crypto/af_alg02.c line 24

Quoted 7 lines
testcases/kernel/crypto/af_alg02.c +5 -1
+#include "tst_safe_pthread.h"
+#include <pthread.h>
+#include <time.h>
+#include <errno.h>
 
-static void run(void)
+#define VERIFY_TIMEOUT (time(NULL) + 5)

It is very neccessary to take some action before process being killed in timeout. In LTP, we have an tst_timeout_remaining() function. Have a look?

Quoted 44 lines
testcases/kernel/crypto/af_alg02.c +25 -0
+
+void *verify_encrypt(void *arg)
 {
        char buf[16];
        int reqfd = tst_alg_setup_reqfd("skcipher", "salsa20", NULL, 16);
@@ -22,6 +33,29 @@ static void run(void)
                tst_res(TPASS, "Successfully \"encrypted\" an empty
 message");
        else
                tst_res(TBROK, "read() didn't return 0");
+       return arg;
+}
+
+static void run(void)
+{
+       pthread_t thr;
+       int join_ret;
+       struct timespec read_timeout;
+
+       read_timeout.tv_sec  = VERIFY_TIMEOUT;
+       read_timeout.tv_nsec = 0;
+
+       SAFE_PTHREAD_CREATE(&thr, NULL, verify_encrypt, NULL);
+       join_ret = pthread_timedjoin_np(thr, NULL, &read_timeout);
+
+       if (join_ret != 0) {
+               if (join_ret == ETIMEDOUT)
+                       tst_brk(TBROK,
+                               "Timed out while reading from request
 socket.");
+               else
+                       tst_brk(TBROK | TTERRNO,
+                               "Error while joining child thread");
+       }
 }

static struct tst_test test = {
--
2.16.4

--
Mailing list info: https://lists.linux.it/listinfo/ltp

Signature · 5 lines
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20190507/91dec0ac/attachment.html>
#2 lore raw

PATCH 1/1 crypto/af_alg02: fixed read() being stuck

Hi Li,

Thanks for having a look at my patch. To my knowledge /pthread_timedjoin_np/ does not kill the thread if it times out. It returns ETIMEDOUT which, in this case, leads to the whole testcase terminating.

I chose this method over setting test.timeout because this way an informative error-message can be printed.

I may be missing the problem here. Can you point me in the right direction why this is bad practice?

Kind regards,

Christian

Quoted 105 lines · Li Wang

Hi Christian,

On Tue, May 7, 2019 at 2:56 PM Christian Amann <[email protected] <mailto:[email protected]>> wrote:

    On kernels < 4.14 (missing commit 2d97591ef43d) reading from
    the request socket does not return and the testcase does not
    finish.

    This fix moves the logic to a child thread in order for the
    parent to handle the timeout and report a message to the user.

    Signed-off-by: Christian Amann <[email protected]
    <mailto:[email protected]>>
    ---
     testcases/kernel/crypto/Makefile   |  2 ++
     testcases/kernel/crypto/af_alg02.c | 36
    +++++++++++++++++++++++++++++++++++-
     2 files changed, 37 insertions(+), 1 deletion(-)

    diff --git a/testcases/kernel/crypto/Makefile
    b/testcases/kernel/crypto/Makefile
    index 76f9308c2..6547e1cb6 100644
    --- a/testcases/kernel/crypto/Makefile
    +++ b/testcases/kernel/crypto/Makefile
    @@ -20,3 +20,5 @@ include $(top_srcdir)/include/mk/testcases.mk
    <http://testcases.mk>
     CFLAGS                 += -D_GNU_SOURCE

     include $(top_srcdir)/include/mk/generic_leaf_target.mk
    <http://generic_leaf_target.mk>
    +
    +af_alg02: CFLAGS += -pthread
    diff --git a/testcases/kernel/crypto/af_alg02.c
    b/testcases/kernel/crypto/af_alg02.c
    index a9e820423..056511993 100644
    --- a/testcases/kernel/crypto/af_alg02.c
    +++ b/testcases/kernel/crypto/af_alg02.c
    @@ -7,12 +7,23 @@
      * Regression test for commit ecaaab564978 ("crypto: salsa20 - fix
      * blkcipher_walk API usage"), or CVE-2017-17805.  This test
    verifies that an
      * empty message can be encrypted with Salsa20 without crashing
    the kernel.
    + *
    + * read() fix:
    + * Calls read() in child thread in order to manually kill it
    after timeout.
    + * With kernels missing commit 2d97591ef43d ("crypto: af_alg -
    consolidation
    + * of duplicate code") read() does not return.
      */

     #include "tst_test.h"
     #include "tst_af_alg.h"
    +#include "tst_safe_pthread.h"
    +#include <pthread.h>
    +#include <time.h>
    +#include <errno.h>

    -static void run(void)
    +#define VERIFY_TIMEOUT (time(NULL) + 5)

It is very neccessary to take some action before process being killed in timeout. In LTP, we have an tst_timeout_remaining() function. Have a look?

    +
    +void *verify_encrypt(void *arg)
     {
            char buf[16];
            int reqfd = tst_alg_setup_reqfd("skcipher", "salsa20",
    NULL, 16);
    @@ -22,6 +33,29 @@ static void run(void)
                    tst_res(TPASS, "Successfully \"encrypted\" an
    empty message");
            else
                    tst_res(TBROK, "read() didn't return 0");
    +       return arg;
    +}
    +
    +static void run(void)
    +{
    +       pthread_t thr;
    +       int join_ret;
    +       struct timespec read_timeout;
    +
    +       read_timeout.tv_sec  = VERIFY_TIMEOUT;
    +       read_timeout.tv_nsec = 0;
    +
    +       SAFE_PTHREAD_CREATE(&thr, NULL, verify_encrypt, NULL);
    +       join_ret = pthread_timedjoin_np(thr, NULL, &read_timeout);
    +
    +       if (join_ret != 0) {
    +               if (join_ret == ETIMEDOUT)
    +                       tst_brk(TBROK,
    +                               "Timed out while reading from
    request socket.");
    +               else
    +                       tst_brk(TBROK | TTERRNO,
    +                               "Error while joining child thread");
    +       }
     }
Quoted 13 lines
     static struct tst_test test = {
    -- 
    2.16.4


    -- 
    Mailing list info: https://lists.linux.it/listinfo/ltp
Signature · 2 lines
Regards,
Li Wang

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20190507/b6480749/attachment.html>

#3 lore raw

PATCH 1/1 crypto/af_alg02: fixed read() being stuck

Quoted 7 lines · Christian Amann

Hi Li,

Thanks for having a look at my patch. To my knowledge *pthread_timedjoin_np* does not kill the thread if it times out. It returns ETIMEDOUT which, in this case, leads to the whole testcase terminating.

Yes, you method could be work but not elegant. AFAIK, pthread_timedjoin_np is non-standard and first appeared in glibc in version 2.3.3

Quoted 6 lines

I chose this method over setting test.timeout because this way an informative error-message can be printed.

I may be missing the problem here. Can you point me in the right direction why this is bad practice?

Seems you misundertood my words:). I was not suggest to set timeout for the test before. As I saw you define VERIFY_TIMEOUT by yourself and involves many variables to revoke thread when it hits reading stuck. That looks a bit redundant.

Maybe we can take use of tst_timeout_remaining() to make things simpler?

we could call:

pthread_setcancelstate(PTHREAD_CANCEL_ENABLE)

at the start of parent, then creat thread to run test for a while of time, until the tst_timeout_remaining() < 10sec, then call:

pthread_cancel(<thread_b_pid>)

At last, print something to declear if necessary.

Signature · 5 lines
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20190507/5e0101ab/attachment.html>