Skip to contentSkip to conversation

cgroup/cgroup_regression_test: Fix umount failure

PATCH Superseded Leo Liang 3 participants 6 messages Jul 16, 2021 – Jul 19, 2021

v3 of 7 Superseded Latest is v7 →

lore
Share & tools
lore

Leo Liang

author
#0 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

The test sequence

	mount -t cgroup -o <controllers> <path>
	mkdir <path>/<dir>
	rmdir <path>/<dir>
	umount <path>
	mount -t cgroup -o <controllers> <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
 .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +5 -5
index 1f7f3820e..056166f11 100755
@@ -145,7 +145,7 @@ test2()
 	fi
 
 	rmdir cgroup/0 cgroup/1
-	umount cgroup/
+	tst_umount cgroup/
 }
 
 #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
 	wait $pid2 2>/dev/null
 
 	rmdir $cpu_subsys_path/0 2> /dev/null
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -222,7 +222,7 @@ test4()
 	mount -t cgroup -o none,name=foo cgroup cgroup/
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/
 
 	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
 		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
 	mount -t cgroup none cgroup 2> /dev/null
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -290,7 +290,7 @@ test6()
 
 	mount -t cgroup -o ns xxx cgroup/ > /dev/null 2>&1
 	rmdir cgroup/[1-9]* > /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/
 	check_kernel_bug
 }
 
Signature · 1 line
2.17.0
#1 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

Hi Leo

Quoted 26 lines

The test sequence

	mount -t cgroup -o<controllers>  <path>
	mkdir<path>/<dir>
	rmdir<path>/<dir>
	umount<path>
	mount -t cgroup -o<controllers>  <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
  .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
index 1f7f3820e..056166f11 100755

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test2() · line 148

Quoted 4 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -1
@@ -145,7 +145,7 @@ test2()
  	fi
 
  	rmdir cgroup/0 cgroup/1
-	umount cgroup/

With adding a comment like
# Avoid possible EBUSY error

so when other converts this case into new api, they will keep to use tst_umount.

Quoted 14 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +2 -1
+	tst_umount cgroup/
  }
 
  #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
  	wait $pid2 2>/dev/null
 
  	rmdir $cpu_subsys_path/0 2>  /dev/null
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test4() · line 225

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -222,7 +222,7 @@ test4()
  	mount -t cgroup -o none,name=foo cgroup cgroup/
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 13 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
  	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
  		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
  	mount -t cgroup none cgroup 2>  /dev/null
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test6() · line 293

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -290,7 +290,7 @@ test6()
 
  	mount -t cgroup -o ns xxx cgroup/>  /dev/null 2>&1
  	rmdir cgroup/[1-9]*>  /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 3 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
  	check_kernel_bug
  }
 
#2 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

Quoted 29 lines

Hi Leo

Quoted 27 lines

The test sequence

	mount -t cgroup -o<controllers>  <path>
	mkdir<path>/<dir>
	rmdir<path>/<dir>
	umount<path>
	mount -t cgroup -o<controllers>  <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
  .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
index 1f7f3820e..056166f11 100755
@@ -145,7 +145,7 @@ test2()
  	fi
Quoted 7 lines

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test2() · line 148

Quoted 2 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -1
  	rmdir cgroup/0 cgroup/1
-	umount cgroup/

With adding a comment like
# Avoid possible EBUSY error

so when other converts this case into new api, they will keep to use tst_umount.

+1

Kind regards,
Petr

Quoted 43 lines
Quoted 14 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +2 -1
+	tst_umount cgroup/
  }
 
  #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
  	wait $pid2 2>/dev/null
 
  	rmdir $cpu_subsys_path/0 2>  /dev/null
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test4() · line 225

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -222,7 +222,7 @@ test4()
  	mount -t cgroup -o none,name=foo cgroup cgroup/
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 12 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
  	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
  		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
  	mount -t cgroup none cgroup 2>  /dev/null
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test6() · line 293

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -290,7 +290,7 @@ test6()
 
  	mount -t cgroup -o ns xxx cgroup/>  /dev/null 2>&1
  	rmdir cgroup/[1-9]*>  /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 2 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
  	check_kernel_bug
  }
author
#3 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

In reply to [email protected]

Hi Yang,

Quoted 29 lines · On Mon, Jul 19, 2021 at 01:24:12PM +0800, [email protected] wrote:

Hi Leo

Quoted 28 lines

The test sequence

	mount -t cgroup -o<controllers>  <path>
	mkdir<path>/<dir>
	rmdir<path>/<dir>
	umount<path>
	mount -t cgroup -o<controllers>  <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
  .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
index 1f7f3820e..056166f11 100755
@@ -145,7 +145,7 @@ test2()
  	fi
 
Quoted 8 lines

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test2() · line 148

Quoted 2 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -1
  	rmdir cgroup/0 cgroup/1
-	umount cgroup/

With adding a comment like
# Avoid possible EBUSY error

so when other converts this case into new api, they will keep to use tst_umount.

No problem! Will do in the next version!
Thanks!

Best regards,
Leo

Quoted 43 lines
Quoted 14 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +2 -1
+	tst_umount cgroup/
  }
 
  #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
  	wait $pid2 2>/dev/null
 
  	rmdir $cpu_subsys_path/0 2>  /dev/null
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test4() · line 225

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -222,7 +222,7 @@ test4()
  	mount -t cgroup -o none,name=foo cgroup cgroup/
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 13 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
  	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
  		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
  	mount -t cgroup none cgroup 2>  /dev/null
  	mkdir cgroup/0
  	rmdir cgroup/0
-	umount cgroup/ 2>  /dev/null
+	tst_umount cgroup/ 2>  /dev/null
  	check_kernel_bug
  }
 

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test6() · line 293

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -290,7 +290,7 @@ test6()
 
  	mount -t cgroup -o ns xxx cgroup/>  /dev/null 2>&1
  	rmdir cgroup/[1-9]*>  /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/

here as well

Quoted 3 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
  	check_kernel_bug
  }
 
#4 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

In reply to Leo Liang
Quoted 26 lines

The test sequence

	mount -t cgroup -o <controllers> <path>
	mkdir <path>/<dir>
	rmdir <path>/<dir>
	umount <path>
	mount -t cgroup -o <controllers> <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
 .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
index 1f7f3820e..056166f11 100755

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test2() · line 148

Quoted 5 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
@@ -145,7 +145,7 @@ test2()
 	fi
 
 	rmdir cgroup/0 cgroup/1
-	umount cgroup/
+	tst_umount cgroup/

As I note in previous patch, with suggested change to tst_umount() to always add trailing "/" it'd be enough just "tst_umount cgroup".

I'm not sure what is better: automatic handling or having to explicitly add trailing "/".

Kind regards,
Petr

Quoted 38 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +4 -4
 }
 
 #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
 	wait $pid2 2>/dev/null
 
 	rmdir $cpu_subsys_path/0 2> /dev/null
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -222,7 +222,7 @@ test4()
 	mount -t cgroup -o none,name=foo cgroup cgroup/
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/
 
 	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
 		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
 	mount -t cgroup none cgroup 2> /dev/null
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -290,7 +290,7 @@ test6()
 
 	mount -t cgroup -o ns xxx cgroup/ > /dev/null 2>&1
 	rmdir cgroup/[1-9]* > /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/
 	check_kernel_bug
 }
author
#5 lore raw

PATCH v3 2/2 cgroup/cgroup_regression_test: Fix umount failure

Hi Petr,

Quoted 28 lines · Petr Vorel
Quoted 27 lines

The test sequence

	mount -t cgroup -o <controllers> <path>
	mkdir <path>/<dir>
	rmdir <path>/<dir>
	umount <path>
	mount -t cgroup -o <controllers> <path>

would easily fail at the last mount with -EBUSY on certain platform.

The reason is that this test sequence would have the chance of missing a release code path when doing rmdir and umount.

Adding a little delay between "rmdir" and "umount" could fix the problem, so use tst_umount API instead of umount in "rmdir, umount" sequence.

Fixes #839
---
 .../controllers/cgroup/cgroup_regression_test.sh       | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +0 -0
index 1f7f3820e..056166f11 100755
@@ -145,7 +145,7 @@ test2()
 	fi
Quoted 6 lines

testcases/kernel/controllers/cgroup/cgroup_regression_test.sh test2() · line 148

Quoted 3 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +1 -1
 	rmdir cgroup/0 cgroup/1
-	umount cgroup/
+	tst_umount cgroup/

As I note in previous patch, with suggested change to tst_umount() to always add trailing "/" it'd be enough just "tst_umount cgroup".

I think your method is better that we don't alter the argument passed into tst_umount. Let users decide the argument and honor the choice.

I will send a v4 patch later and still keep the trailing slash.

Thanks!

Best regards,
Leo

Quoted 44 lines

I'm not sure what is better: automatic handling or having to explicitly add trailing "/".

Kind regards,
Petr

Quoted 38 lines
testcases/kernel/controllers/cgroup/cgroup_regression_test.sh +4 -4
 }
 
 #---------------------------------------------------------------------------
@@ -193,7 +193,7 @@ test3()
 	wait $pid2 2>/dev/null
 
 	rmdir $cpu_subsys_path/0 2> /dev/null
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -222,7 +222,7 @@ test4()
 	mount -t cgroup -o none,name=foo cgroup cgroup/
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/
+	tst_umount cgroup/
 
 	if dmesg | grep -q "MAX_LOCKDEP_SUBCLASSES too low"; then
 		tst_res TFAIL "lockdep BUG was found"
@@ -254,7 +254,7 @@ test5()
 	mount -t cgroup none cgroup 2> /dev/null
 	mkdir cgroup/0
 	rmdir cgroup/0
-	umount cgroup/ 2> /dev/null
+	tst_umount cgroup/ 2> /dev/null
 	check_kernel_bug
 }
 
@@ -290,7 +290,7 @@ test6()
 
 	mount -t cgroup -o ns xxx cgroup/ > /dev/null 2>&1
 	rmdir cgroup/[1-9]* > /dev/null 2>&1
-	umount cgroup/
+	tst_umount cgroup/
 	check_kernel_bug
 }