semctl01.c: Pass the correct parameter

PATCH Superseded DAN LI 2 participants 3 messages May 21, 2013 – May 22, 2013

Share & tools

DAN LI

author

2013-05-21 08:54 UTC #0 lore raw

PATCH 1/1 semctl01.c: Pass the correct parameter

For SEM_STAT, the semid argument is not a semaphore identifier, but instead an index into the kernel’s internal array that maintains information about all semaphore sets on the system.

Pass a correct index of the kernel' internal array intead of a semaphores id when testing feature SEM_STAT.

Signed-off-by DAN LI <[email protected]>

---
 testcases/kernel/syscalls/ipc/semctl/semctl01.c | 28 +++++++++++++++++++------
 1 file changed, 22 insertions(+), 6 deletions(-)

testcases/kernel/syscalls/ipc/semctl/semctl01.c +22 -6

index 687d909..aee6ccf 100644
@@ -52,6 +52,8 @@ char *TCID = "semctl01";
 int TST_TOTAL = 13;
 
 static int sem_id_1 = -1;
+static int sem_id;
+static int sem_index;
 
 static int sync_pipes[2];

@@ -67,12 +69,12 @@ static void pid_setup(void), func_pid(int);
 static void func_gval(int);
 static void sall_setup(void), func_sall(void);
 static void func_sval(void);
-static void func_rmid(void);
+static void rmid_setup(void), func_rmid(void);
 static void child_cnt(void);
 static void child_pid(void);
 static void func_iinfo(int);
 static void func_sinfo(void);
-static void func_sstat(int);
+static void sstat_setup(void), func_sstat(int);
 
 static struct semid_ds buf;
 static struct seminfo ipc_buf;

@@ -117,8 +119,8 @@ static struct test_case_t {
 	{SEM4, SETVAL, func_sval, SEMUN_CAST INCVAL, NULL},
 	{0, IPC_INFO, func_iinfo, SEMUN_CAST & ipc_buf, NULL},
 	{0, SEM_INFO, func_sinfo, SEMUN_CAST & ipc_buf, NULL},
-	{0, SEM_STAT, func_sstat, SEMUN_CAST & buf, NULL},
-	{0, IPC_RMID, func_rmid, SEMUN_CAST & buf, NULL},
+	{0, SEM_STAT, func_sstat, SEMUN_CAST & buf, sstat_setup},
+	{0, IPC_RMID, func_rmid, SEMUN_CAST & buf, rmid_setup},
 };
 
 int main(int argc, char *argv[])

@@ -538,6 +540,11 @@ static void func_sval(void)
 		tst_resm(TPASS, "semaphore value is correct");
 }
 
+static void rmid_setup(void)
+{
+	sem_id_1 = sem_id;
+}
+
 /*
  * func_rmid() - check the functionality of the IPC_RMID command with semctl()
  */

@@ -560,10 +567,13 @@ static void func_rmid(void)
 
 static void func_iinfo(int hidx)
 {
-	if (hidx >= 0)
+	if (hidx >= 0) {
+		sem_index = hidx;
 		tst_resm(TPASS, "the highest index is correct");
-	else
+	} else {
+		sem_index = 0;
 		tst_resm(TFAIL, "the highest index is incorrect");
+	}
 }
 
 static void func_sinfo(void)

@@ -574,6 +584,12 @@ static void func_sinfo(void)
 		tst_resm(TPASS, "number of semaphore sets is correct");
 }
 
+static void sstat_setup(void)
+{
+	sem_id = sem_id_1;
+	sem_id_1 = sem_index;
+}
+
 static void func_sstat(int semidx)
 {
 	if (semidx >= 0)

Signature · 12 lines

1.7.11.rc0

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Jan Stancek

2013-05-21 10:22 UTC #1 lore raw

----- Original Message -----

Quoted 11 lines

From: "DAN LI" <[email protected]>
To: "LTP list" <[email protected]>
Sent: Tuesday, 21 May, 2013 10:51:53 AM
Subject: [LTP] [PATCH] semctl01.c: Pass the correct parameter
For SEM_STAT, the semid argument is not a semaphore identifier, but instead an
index into the kernel’s internal array that maintains information about all semaphore sets on the system.

Quoted 6 lines

Pass a correct index of the kernel' internal array intead of a semaphores id when
testing feature SEM_STAT.
Signed-off-by DAN LI <[email protected]>
---

Hi,

that backup/restore of global variable in setup/cleanup funcs between testcases looks a bit odd to me.

How about adding new field to test_case_t?

+static int sem_index = 0;
static struct test_case_t {
+ int *semid;
...
- TEST(semctl(sem_id_1, TC[i].semnum, TC[i].cmd,
+ TEST(semctl(*(TC[i].semid), TC[i].semnum, TC[i].cmd,

These you could initialise in TC[] definition to either &sem_id_1 or &sem_index and avoid all that backup/restore of sem_id_1.

What do you think?

Regards,
Jan

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, & servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

DAN LI

author

2013-05-22 01:50 UTC #2 lore raw

Quoted 37 lines · Jan Stancek

----- Original Message -----
Quoted 11 lines
From: "DAN LI" <[email protected]>
To: "LTP list" <[email protected]>
Sent: Tuesday, 21 May, 2013 10:51:53 AM
Subject: [LTP] [PATCH] semctl01.c: Pass the correct parameter
For SEM_STAT, the semid argument is not a semaphore identifier, but instead an
index into the kernel’s internal array that maintains information about all semaphore sets on the system.
Quoted 6 lines
Pass a correct index of the kernel' internal array intead of a semaphores id when
testing feature SEM_STAT.
Signed-off-by DAN LI <[email protected]>
---
Hi,
that backup/restore of global variable in setup/cleanup funcs between testcases looks a bit odd to me.
How about adding new field to test_case_t?
(quoted diff) +3 -1
+static int sem_index = 0;
 static struct test_case_t {
+       int *semid;
...
-                       TEST(semctl(sem_id_1, TC[i].semnum, TC[i].cmd,
+                       TEST(semctl(*(TC[i].semid), TC[i].semnum, TC[i].cmd,

Quoted 4 lines

These you could initialise in TC[] definition to either &sem_id_1 or &sem_index and avoid all that backup/restore of sem_id_1.
What do you think?

Sounds better.;)

Thank you for reviewing.

Regards,
DAN LI

Quoted 4 lines

Regards,
Jan