Eric W. Biederman wrote:
[ ... ]
I guess my meaning is I was expecting.
child = fork();
if (child == 0) {
execve(...);
}
waitpid(child);
This puts /bin/sh in the container as well.
#include unistd.h
#include stdlib.h
#include stdio.h
#include syscall.h
#include sys/types.h
Daniel Lezcano daniel.lezc...@free.fr writes:
Eric W. Biederman wrote:
[ ... ]
I guess my meaning is I was expecting.
child = fork();
if (child == 0) {
execve(...);
}
waitpid(child);
This puts /bin/sh in the container as well.
#include unistd.h
#include stdlib.h
#include
Eric W. Biederman wrote:
Daniel Lezcano daniel.lezc...@free.fr writes:
Eric W. Biederman wrote:
[ ... ]
I guess my meaning is I was expecting.
child = fork();
if (child == 0) {
execve(...);
}
waitpid(child);
This puts /bin/sh in the container as well.
When we restart an application, we won't try to restore the original
pid in the parent pid_ns, so we don't checkpoint that pid.
However, if we are going to dump mount info from userspace using
/proc/pid/mountinfo, then it will be easiest for a restart wrapper
to use the pids from the checkpoint
This will be used by a restart wrapper to analyze /proc/$$/mountinfo.
Also add the new rpids field to ckpt_pids to reflect kernel.
It depends on the patch I just sent for linux-cr adding rpid to
struct ckpt_pids.
Signed-off-by: Serge E. Hallyn se...@us.ibm.com
---
ckptinfo.c
This outputs a bunch of garbage but according to the comments don't
absolutely need this wait anyway. So redirect stderr for the bash
wait to /dev/null.
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
counterloop/cloop_serial.sh |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
rewrite-cr-header.sh |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/rewrite-cr-header.sh b/rewrite-cr-header.sh
index 62ad5b5..21f699c 100755
--- a/rewrite-cr-header.sh
+++ b/rewrite-cr-header.sh
@@ -45,6 +45,7 @@ set
do_ckpt() sets checkpoint-ready and waits for checkpoint-done.
Test scripts should set checkpoint-skip if tests should skip
the wait for checkpoint-done as if checkpointing were done.
This allows us to have tests with multiple checkpoints in them:
checkpoint 1
...
The perror() after calling checkpoint() is useless -- stderr is closed.
Fix it by using dup2() to redirect stderr to the specified file. Do
the same for stdout just in case we want to use printf later.
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
simple/ckpt.c |4 ++--
1 files
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
eventfd/libeptest.h | 46 +-
1 files changed, 1 insertions(+), 45 deletions(-)
diff --git a/eventfd/libeptest.h b/eventfd/libeptest.h
index b6629e8..3f1522e 100644
--- a/eventfd/libeptest.h
+++
I'd have thought seeing hello, world! was a sign the test passed. Make the
output indicative of the test results.
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
simple/ckpt.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/simple/ckpt.c b/simple/ckpt.c
index
cr.h is grossly out-of-date. Any build which relies on it is horribly broken
without this change.
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
cr.h | 46 +++---
1 files changed, 19 insertions(+), 27 deletions(-)
diff --git a/cr.h b/cr.h
index
These weren't being cleaned up.
Signed-off-by: Matt Helsley matth...@us.ibm.com
---
epoll/module.mk |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/epoll/module.mk b/epoll/module.mk
index 2e5d8c6..260e549 100644
--- a/epoll/module.mk
+++ b/epoll/module.mk
@@ -9,3 +9,5
Quoting Matt Helsley (matth...@us.ibm.com):
On Tue, Mar 09, 2010 at 03:04:51PM -0600, Serge E. Hallyn wrote:
When we restart an application, we won't try to restore the original
pid in the parent pid_ns, so we don't checkpoint that pid.
However, if we are going to dump mount info from
On Tue, Mar 09, 2010 at 03:04:51PM -0600, Serge E. Hallyn wrote:
When we restart an application, we won't try to restore the original
pid in the parent pid_ns, so we don't checkpoint that pid.
However, if we are going to dump mount info from userspace using
/proc/pid/mountinfo, then it will
Quoting Matt Helsley (matth...@us.ibm.com):
This outputs a bunch of garbage but according to the comments don't
absolutely need this wait anyway. So redirect stderr for the bash
wait to /dev/null.
Signed-off-by: Matt Helsley matth...@us.ibm.com
All applied, thanks.
-serge
* Andrea Righi ari...@develer.com [2010-03-10 00:00:31]:
Control the maximum amount of dirty pages a cgroup can have at any given time.
Per cgroup dirty limit is like fixing the max amount of dirty (hard to
reclaim)
page cache used by any cgroup. So, in case of multiple cgroup writers,
Please please measure the performance overhead of this change.
here.
I made a patch below and measured the time(average of 10 times) of
kernel build
on tmpfs(make -j8 on 8 CPU machine with 2.6.33 defconfig).
before
- root cgroup: 190.47 sec
- child
* nishim...@mxp.nes.nec.co.jp nishim...@mxp.nes.nec.co.jp [2010-03-10
10:43:09]:
Please please measure the performance overhead of this change.
here.
I made a patch below and measured the time(average of 10 times)
of kernel build
on tmpfs(make -j8 on 8 CPU machine
19 matches
Mail list logo