n
perhaps ?
Please note that we also have _per_ user quota accounting, the trade off
of accouting prevents further performance penalties on other bus
operations. Referring to performance critical code, this code path can
just be ignored by to not opt-in for KDBUS_ATTACH_CAPS which is the
default
uot;,
> > +[SERVICE_FAILURE_TIMEOUT] = "timeout",
> > +[SERVICE_FAILURE_EXIT_CODE] = "exit-code",
> > +[SERVICE_FAILURE_SIGNAL] = "signal",
> > +[SERVICE_FAILURE_CORE_DUMP] = "core-dump",
> > +[SER
Hi Shawn,
On Tue, Mar 31, 2015 at 04:59:29PM -0700, Shawn Landden wrote:
> On Tue, Mar 31, 2015 at 2:40 PM, Djalal Harouni wrote:
> > On Tue, Mar 31, 2015 at 11:10:34AM -0700, Shawn Landden wrote:
[...]
> >> The point is that assert() and assert_se() should only be used fo
On Tue, Mar 31, 2015 at 11:10:34AM -0700, Shawn Landden wrote:
> On Tue, Mar 31, 2015 at 8:38 AM, Djalal Harouni wrote:
> > On Mon, Mar 30, 2015 at 07:32:35PM -0700, Shawn Landden wrote:
> >> On Mon, Mar 30, 2015 at 5:04 PM, Djalal Harouni wrote:
> >> > On Fri, Ma
On Mon, Mar 30, 2015 at 07:32:35PM -0700, Shawn Landden wrote:
> On Mon, Mar 30, 2015 at 5:04 PM, Djalal Harouni wrote:
> > On Fri, Mar 27, 2015 at 09:51:26AM -0700, Shawn Landden wrote:
> >> On Fri, Mar 27, 2015 at 8:16 AM, Tom Gundersen wrote:
> > [...]
> >> &
s. It is not to make it
> faster, I think that is negligible.
Hmm embedded cases are real, I had to deal with some in the past. But
not sure here since I didn't see any numbers before/after stripping, but
perhaps you can start by updating the callers and their semantics if y
On Fri, Mar 27, 2015 at 04:16:55PM +0100, Tom Gundersen wrote:
> On Fri, Mar 27, 2015 at 2:04 PM, Djalal Harouni wrote:
> > Hi Shawn,
> >
> > On Thu, Mar 26, 2015 at 11:21:54PM -0700, Shawn Landden wrote:
> >> On Thu, Mar 26, 2015 at 5:47 PM, Djalal Harouni wrote:
Hi Shawn,
On Thu, Mar 26, 2015 at 11:21:54PM -0700, Shawn Landden wrote:
> On Thu, Mar 26, 2015 at 5:47 PM, Djalal Harouni wrote:
> > On Fri, Mar 27, 2015 at 12:30:53AM +0100, Tom Gundersen wrote:
> >> On Thu, Mar 26, 2015 at 9:19 AM, Lennart Poettering
> >> wrote:
&g
ndif
> >>
> >> --
> >> 2.2.1.209.g41e5f3a
> >>
> >> ___
> >> systemd-devel mailing list
> >> systemd-devel@lists.freedesktop.org
> >> http://lists.freedesktop.org/mailman/listinfo/systemd-
forget about them even if they are
called later from libraries or copy+past into a library call... ?!
> Thanks
> David
> ___
> systemd-devel mailing list
> systemd-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/list
On Thu, Feb 19, 2015 at 05:44:34PM +0100, Djalal Harouni wrote:
> On Thu, Feb 19, 2015 at 01:05:22PM +, Simon McVittie wrote:
> > On 19/02/15 12:43, Lukasz Skalski wrote:
> > >GetConnectionCredentials method was added to dbus-1 specification
> > >more than one y
gh kdbus knowledge to
> know about those might as well be using kdbus directly.
>
> >+r = get_creds_by_message(a, m,
> >SD_BUS_CREDS_PID|SD_BUS_CREDS_EUID, &creds, &error);
>
> Can this ever return "unknown" (-1?) for creds->pid or creds->e
USER16
> +#define KDBUS_CONN_MAX_MSGS_PER_USER128
>
> /**
> * Note: this test can be used to inspect policy_db->talk_access_hash
> --
> 1.9.3
>
> ___
> systemd-devel mailing list
> systemd-devel@li
Add missing IFLA_INET6_ADDR_GEN_MODE definition so we can build with
kernel headers < 3.17
---
configure.ac | 3 ++-
src/shared/missing.h | 18 ++
2 files changed, 20 insertions(+), 1 deletion(-)
diff --git a/configure.ac b/configure.ac
index 684c793..97a29d6 100644
--- a
On Wed, Feb 11, 2015 at 05:06:56PM +0100, Lennart Poettering wrote:
> On Wed, 11.02.15 13:53, Djalal Harouni (tix...@opendz.org) wrote:
>
> > On Tue, Feb 10, 2015 at 12:52:34PM +0100, Lennart Poettering wrote:
> > > On Thu, 05.02.15 02:03, Vasiliy Tolstov (v.tol
his is not true,
we don't have CAP_SYS_MODULE... this will confuse programs running
inside containers, we'll have to add more code paths for this special
case... and not only CAP_SYS_MODULE, perhaps there are other cases...
--
Djalal Harouni
http://opendz.org
On Mon, Jan 19, 2015 at 02:42:31PM +0200, Mantas Mikulėnas wrote:
> On Mon, Jan 19, 2015 at 2:26 PM, Djalal Harouni wrote:
>
> > Hi,
> >
> > On Mon, Jan 19, 2015 at 10:46:23AM +0100, Rémi Audebert wrote:
> > > Signed-off-by: Rémi Audebert
> > Your email
to be sent. An attempt to send less that that should result
> - * in -ECONNREFUSED.
> + * to be sent. An attempt to send less than that should result in
> + * -ECONNREFUSED.
>*/
> hello.attach_flags_send = _KDBUS_ATTACH_ALL & ~KDBUS_ATTACH_TIMESTAMP;
space,
major ones do not support this...
> Tom
>
> [0]:
> [1]:
>
> [root@tomegun-x240 userns]# ./userns_child_exec -U -M '0 0 4294967295'
> -G '0 0 4294967295' bash
> [root@tomegun-x240 userns]# mknod null b 1 3
> mknod: ‘null’: Operation not perm
eparate user namespace and still
> have the same map as the host is if you had a privileged helper write
> the whole host uid and gid map to your process' uid_map and gid_map, in
> which case, your process' uid 0 is mapped to the host uid 0 and while
> you are technically in a different namespace than the host, yo
> MS_RDONLY : 0), NULL) < 0) {
> > +r = get_mount_flags(x, &orig_flags);
> > +if (r < 0)
> > + return r;
> > +orig_flags &= ~MS_RDONLY;
> > +
/shared/util.c
> @@ -61,6 +61,7 @@
> #include
> #include
> #include
> +#include
> #undef basename
>
> #ifdef HAVE_SYS_AUXV_H
> @@ -6858,6 +6859,15 @@ int umount_recursive(const char *prefix, int flags) {
> return r ? r : n;
> }
>
> +uns
custom endpoints
/sys/fs/kdbus// [1], the whole thing is mounted on top
of tmpfs in order to hide the rest of the kdbufs tree
The custom endpoints are created when creating services, this will be
part of the namespaced apps setup where apps will only see a subset of
names on a bus... the cus
if they cannot be translated due to namespaces.
> >
> > Thanks. Could you clarify in the docs?
>
> Daniel, David? Could you add a note about this?
Ok pushed a note about namespace issues, thanks!
--
Djalal Harouni
http://opendz.org
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/systemd-devel
ged (capabilities gathered when the message was queued).
Hope I'm not missing something, otherwise let me know, thank you Simon!
> S
>
> ___
> systemd-devel mailing list
> systemd-dev
Signed-off-by: Djalal Harouni
---
test/kdbus-util.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/test/kdbus-util.c b/test/kdbus-util.c
index fe4565c..b1c5864 100644
--- a/test/kdbus-util.c
+++ b/test/kdbus-util.c
@@ -569,7 +569,7 @@ void kdbus_msg_dump(const struct
best practice and to be able to free the returned
kdbus_msg and the slice pointed by that offset.
Signed-off-by: Djalal Harouni
---
Hi Daniel, before applying please make sure that we want this. It
follows what I've discussed in the other mail, otherwise just test it,
it will give a better vi
On Tue, Sep 09, 2014 at 10:40:57AM +0200, Daniel Mack wrote:
> On 09/08/2014 03:50 PM, Djalal Harouni wrote:
> > Yes there are compile time checks, and it is perhaps easier/consistent
> > to read this way! but yes a union is also good. OK I'll update it.
>
> Nevermind
On Mon, Sep 08, 2014 at 03:32:21PM +0200, Daniel Mack wrote:
> On 09/08/2014 03:18 PM, Djalal Harouni wrote:
> > Upstream kernels allow unprivileged users to create user namespaces
> > and change their uid/gid.
> >
> > These patches update kdbus policy logic to handl
On Mon, Sep 08, 2014 at 03:27:42PM +0200, Daniel Mack wrote:
> On 09/08/2014 03:18 PM, Djalal Harouni wrote:
> > * This is the internal version of struct kdbus_policy_db_access.
> > @@ -51,7 +52,8 @@ struct kdbus_policy_db_cache_entry {
> > struct kdbus_policy_db_entry_acc
we can add it later a la:
fs/inode.c:inode_owner_or_capable()
Add kdbus_policy_make_access() to convert the user provided info to the
current user namespace. Userspace struct is not changed, only the kernel
one.
This patch fixes test #8 of test-kdbus-policy
Signed-off-by: Djalal Harouni
.
* kdbus_clone_userns_test() that performs the test inside a new
user namespace.
* Converts all the other tests to return CHECK_OK, CHECK_SKIP or
CHECK_ERR so we are consistent.
Currently we fail at kdbus_clone_userns_test() test #8. The next patch
will fix this issue.
Signed-off-by: Djalal Harouni
---
test
---
Currently this bus_kernel_create_monitor() is not used.
Patch compile tested.
src/libsystemd/sd-bus/bus-kernel.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/libsystemd/sd-bus/bus-kernel.c
b/src/libsystemd/sd-bus/bus-kernel.c
index 3ca271c..1440e43 100644
--- a/sr
On Wed, Aug 20, 2014 at 10:49:22PM +0200, Daniel Mack wrote:
> On 08/20/2014 06:16 PM, Djalal Harouni wrote:
> > On Tue, Aug 19, 2014 at 09:15:35AM +0200, Daniel Mack wrote:
>
> >> Hmm, I'm not convinced this buys us anything really. After all, that
> >>
Signed-off-by: Djalal Harouni
---
test/kdbus-util.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/test/kdbus-util.c b/test/kdbus-util.c
index f79d7ec..5b3df7d 100644
--- a/test/kdbus-util.c
+++ b/test/kdbus-util.c
@@ -471,8 +471,8 @@ void msg_dump(const struct conn
On Wed, Aug 20, 2014 at 06:04:09PM +0200, Daniel Mack wrote:
> On 08/20/2014 05:58 PM, Djalal Harouni wrote:
> > case KDBUS_ITEM_AUDIT:
> > - printf(" +%s (%llu bytes) loginuid=%llu
> > sessionid=%llu\n",
> > +
On Tue, Aug 19, 2014 at 09:15:35AM +0200, Daniel Mack wrote:
> Hi Djalal,
Thanks for applying the others.
> On 08/19/2014 03:43 AM, Djalal Harouni wrote:
> > Some creds can be gathered during kdbus_meta_append() instead of
> > kdbus_conn_queue_alloc() where they will be ga
Do the namespace translation just before pushing into the receiver's
slice, so we map to the receiver's user namespace.
Patch tested.
Signed-off-by: Djalal Harouni
---
connection.c | 37 +
metadata.c | 3 +++
metadata.h | 4
3 files c
We have sessionid then loginuid in kdbus_audit.
Signed-off-by: Djalal Harouni
---
test/kdbus-util.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/test/kdbus-util.c b/test/kdbus-util.c
index f79d7ec..956fa6f 100644
--- a/test/kdbus-util.c
+++ b/test/kdbus-util.c
@@ -469,7
, name, MFD_ALLOW_SEALING);
>
> Ok, cpp is really confused about which headers to look at now. I guess
> we have to rely on the kernel headers being available under
> $(KERNELDIR)/usr/include.
>
> Could you please try if temporarily reverting my top-m
was first reported by Greg, and to compile and test kdbus, I just
added "export KERNELDIR" to the root Makefile, didn't bother to search...
--
Djalal Harouni
http://opendz.org
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/systemd-devel
it.
Patch tested with the test-kdbus-metadata-ns
Signed-off-by: Djalal Harouni
---
connection.c | 56
metadata.c | 6 +-
metadata.h | 26 ++
3 files changed, 67 insertions(+), 21 deletions(-)
diff --git a
Move the install creds into the receiver's slice to its own function
kdbus_conn_creds_install().
Use from_kuid_munged(), so the uid mapping never fails.
Signed-off-by: Djalal Harouni
---
connection.c | 32
1 file changed, 20 insertions(+), 12 deletions(-)
Signed-off-by: Djalal Harouni
---
metadata.c | 3 ++-
metadata.h | 8 +++-
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/metadata.c b/metadata.c
index eb286aa..3dff9ac 100644
--- a/metadata.c
+++ b/metadata.c
@@ -130,7 +130,8 @@ kdbus_meta_append_item(struct kdbus_meta *meta
Hi,
Patch 1 and 2 should be applied.
For patch 3 it can be discussed :-)
Please apply on top of the test series that contain the
test-kdbus-metadta-ns tool.
Thanks!
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freede
ered a compile-time
> constant. I don't know whether conditions with __builtin_constant_p()
> are evaluated at the parser-step. The GCC example replaces the ELSE
> case with -1, effectively making both compile-time constants.
Sorry I didn't follow the thread, but:
Actually and IIRC
all the tests. This
prevents messing up with the attach-flags.
Signed-off-by: Djalal Harouni
---
test/kdbus-util.c| 62 +++-
test/kdbus-util.h| 7 +++---
test/test-kdbus-policy.c | 2 +-
3 files changed, 45 insertions(+), 26 deletions
in it.
And update a kdbus_policy_set() call to only pass a one name per
policy-holding connection
Signed-off-by: Djalal Harouni
---
connection.c | 21 -
1 file changed, 16 insertions(+), 5 deletions(-)
diff --git a/connection.c b/connection.c
index c2d71a8..342c870 100644
Allow KDBUS_CMD_CONN_UPDATE for KDBUS_CONN_POLICY_HOLDER connections.
Signed-off-by: Djalal Harouni
---
handle.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/handle.c b/handle.c
index 2e6502b..ac68681 100644
--- a/handle.c
+++ b/handle.c
@@ -636,8 +636,8 @@ static
Hi,
This goes on top of the previous one:
http://lists.freedesktop.org/archives/systemd-devel/2014-August/021747.html
Kay, sorry it should be perhaps just be a one series, but I just noticed
those bugs, so just send quick fixes.
If you want me to resend as a one series, I will do it, no problem.
If userspace calls in with the wrong connection type, just return
-EOPNOTSUPP instead of -EPERM.
This will not confuse unprivileged and privileged processes, and permits
to identify legitimate -EPERM errors.
This just converts errors introduced in commit 7015a1e6746
Signed-off-by: Djalal
.
Signed-off-by: Djalal Harouni
---
test/test-kdbus-policy.c | 25 +++--
1 file changed, 19 insertions(+), 6 deletions(-)
diff --git a/test/test-kdbus-policy.c b/test/test-kdbus-policy.c
index a2430f2..e0bd619 100644
--- a/test/test-kdbus-policy.c
+++ b/test/test-kdbus
nd in different places of the code we only check for
activators when doing send/recv validation.
So I guess we should also block policy holders from owning well-known
names ? hmm, then add the policy holders to the "block X connections
from sending or receiving
Hi,
On Thu, Jul 31, 2014 at 10:38:47PM +0200, Kay Sievers wrote:
> On Thu, Jul 31, 2014 at 8:57 PM, Djalal Harouni wrote:
> > (Cc'ed Lennart)
> >
> > On Thu, Jul 31, 2014 at 05:40:53PM +0200, Kay Sievers wrote:
> >> On Wed, Jul 23, 2014 at 6:34 PM, Djalal Haro
(Cc'ed Lennart)
On Thu, Jul 31, 2014 at 05:40:53PM +0200, Kay Sievers wrote:
> On Wed, Jul 23, 2014 at 6:34 PM, Djalal Harouni wrote:
> > This series adds the infrastructure to test and upload multiple
> > policies.
> >
> > The last #5 patch allows to upload
Signed-off-by: Djalal Harouni
---
domain.c | 67
domain.h | 2 --
2 files changed, 69 deletions(-)
diff --git a/domain.c b/domain.c
index 1e2c3c9..eeb73ca 100644
--- a/domain.c
+++ b/domain.c
@@ -557,73 +557,6 @@ int
First use kzalloc to allocate the users array, so we do not reference
unintialized values.
And free the old conn->msg_users array not the newly allocated 'users'
one.
Patch tested, and users will hit the KDBUS_CONN_MAX_MSGS_PER_USER limit
and fail with -ENOBUFS
Signed-off-by: D
mes.
kdbus_bus_new():
=> take domain lock
=> check if domain is still active/connected
=> __kdbus_domain_user_account()
...
Signed-off-by: Djalal Harouni
---
bus.c | 13 +
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/bus.c b/bus.c
index d09e3c6..9e
last one in bus.c is updated in the
next patch.
Signed-off-by: Djalal Harouni
---
connection.c | 12 ++--
handle.c | 8
2 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/connection.c b/connection.c
index 1658a92..8838029 100644
--- a/connection.c
+++ b
Add kdbus_domain_user_account() to account and link users into a
domain.
This function will take the domain lock, and it will be used as a
replacement for kdbus_domain_user_find_or_new().
Signed-off-by: Djalal Harouni
---
domain.c | 36
domain.h | 4
Add __kdbus_domain_user_account() to account and link users into a
domain.
Signed-off-by: Djalal Harouni
---
domain.c | 70
domain.h | 4
2 files changed, 74 insertions(+)
diff --git a/domain.c b/domain.c
index af9d986
Signed-off-by: Djalal Harouni
---
domain.c | 33 +
1 file changed, 33 insertions(+)
diff --git a/domain.c b/domain.c
index c4912fa..af9d986 100644
--- a/domain.c
+++ b/domain.c
@@ -419,6 +419,39 @@ int kdbus_domain_make_user(struct kdbus_cmd_make *cmd,
char
Hi,
This is series v2 of:
http://lists.freedesktop.org/archives/systemd-devel/2014-July/021526.html
This series improves user domain accounting and fixes some bugs.
It should go on top of the "kdbus: allow multiple policies" series:
http://lists.freedesktop.org/archives/systemd-devel/2014-July/0
On Wed, Jul 23, 2014 at 10:19:04PM +0100, Djalal Harouni wrote:
> Hi,
>
> This series improves user domain accounting and fixes some bugs.
>
> On top of the "kdbus: allow multiple policies" series:
> http://lists.freedesktop.org/archives/systemd-devel/2014-July/0215
Hi,
It seems there is still another bug in the use/caculation of
conn->msg_users_max, will send another patch on top of this.
On Wed, Jul 23, 2014 at 10:19:11PM +0100, Djalal Harouni wrote:
> First use kzalloc to allocate the users array, so we do not reference
> unintialized values.
First use kzalloc to allocate the users array, so we do not reference
unintialized values.
And free the old conn->msg_users array not the newly allocated 'users'
one.
Patch tested, and users will hit the KDBUS_CONN_MAX_MSGS_PER_USER limit
and fail with -ENOBUFS
Signed-off-by: D
ain lock 2 times by
replacing the kdbus_domain_user_find_or_new() with
kdbus_domain_user_find(): take the lock 1 time
kdbus_domain_user_new()
+
kdbus_bus_new(): take the lock 1 time and use the unlocked version
__kdbus_domain_user_account() to account the user.
Signed-off-by: Djalal Harouni
---
bu
)
So convert the former to 1, and adapt other parts of the code to treat 1
as the starting id of user indexes.
Signed-off-by: Djalal Harouni
---
bus.c| 4 ++--
connection.c | 19 +++
domain.c | 76 +---
handle.c
Add kdbus_domain_user_find() to look up domain users
Signed-off-by: Djalal Harouni
---
domain.c | 32
1 file changed, 32 insertions(+)
diff --git a/domain.c b/domain.c
index 18dc2a7..a5abb2d 100644
--- a/domain.c
+++ b/domain.c
@@ -446,6 +446,38
Add kdbus_domain_user_new() to allocate kdbus_domain_user objects.
Signed-off-by: Djalal Harouni
---
domain.c | 27 +++
1 file changed, 27 insertions(+)
diff --git a/domain.c b/domain.c
index 86fde55..18dc2a7 100644
--- a/domain.c
+++ b/domain.c
@@ -419,6 +419,33 @@ int
Add the lock protected version of __kdbus_domain_user_account(). It
will check if the domain is still active before linking users.
Signed-off-by: Djalal Harouni
---
domain.c | 24
1 file changed, 24 insertions(+)
diff --git a/domain.c b/domain.c
index a321f31..86fde55
Add __kdbus_domain_user_account() to account and link users into a
domain.
Signed-off-by: Djalal Harouni
---
domain.c | 32
1 file changed, 32 insertions(+)
diff --git a/domain.c b/domain.c
index c4912fa..a321f31 100644
--- a/domain.c
+++ b/domain.c
@@ -419,6
Hi,
This series improves user domain accounting and fixes some bugs.
On top of the "kdbus: allow multiple policies" series:
http://lists.freedesktop.org/archives/systemd-devel/2014-July/021514.html
Patches 1, 2, 3 and 4 are preparation patches to improve the code.
Patch 5 fixes kdbus_domain_us
Signed-off-by: Djalal Harouni
---
connection.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/connection.c b/connection.c
index 85ffa5a..1658a92 100644
--- a/connection.c
+++ b/connection.c
@@ -1905,7 +1905,11 @@ int kdbus_conn_new(struct kdbus_ep *ep
Update the policy test in order to register multiple policies
Signed-off-by: Djalal Harouni
---
test/test-kdbus-policy.c | 57 +---
1 file changed, 40 insertions(+), 17 deletions(-)
diff --git a/test/test-kdbus-policy.c b/test/test-kdbus-policy.c
icies size
kdbus_policy_make_item_name() to make a policy item composed of:
KDBUS_ITEM_NAME + (num_access * KDBUS_ITEM_POLICY_ACCESS)
kdbus_policy_make_entries() a helper to construct multiple policies
Signed-off-by: Djalal Harouni
---
test/kdbus-u
Signed-off-by: Djalal Harouni
---
test/kdbus-util.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/test/kdbus-util.h b/test/kdbus-util.h
index 0fcfb72..39d7bb5 100644
--- a/test/kdbus-util.h
+++ b/test/kdbus-util.h
@@ -36,6 +36,12 @@ struct conn {
size_t size;
};
+struct
Signed-off-by: Djalal Harouni
---
test/test-kdbus-policy.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/test/test-kdbus-policy.c b/test/test-kdbus-policy.c
index 6099087..de725e9 100644
--- a/test/test-kdbus-policy.c
+++ b/test/test-kdbus-policy.c
@@ -200,7 +200,7
This series adds the infrastructure to test and upload multiple
policies.
The last #5 patch allows to upload multiple policies per connection
The todo for the policy holders is:
* Should we set a maximum value for how many names/policies a policy
holder is allowed to upload. This is needed sin
On Thu, Jul 17, 2014 at 11:30:26AM +0200, David Herrmann wrote:
> Hi
>
> On Mon, Jul 14, 2014 at 3:28 AM, Djalal Harouni wrote:
> > ppoll is atomic and it is handled by the kernel, so perhaps
> > setting/restoring sigmask can be done easily! and for nspawn: IMO we need
&
EXIT_FAILURE);
> }
>
> +barrier_set_role(&barrier, BARRIER_PARENT);
> fdset_free(fds);
> fds = NULL;
>
> -/* Wait for the child event:
> - * If EVENTFD_CHILD_FAILED, the c
is wrong, your target here is shutdown... and your
process should only do small operations to record something or clean
things up...
So there is the 'PrepareForShutdown' signal, assume SessionRemoved, then
release the inhibitor lock ?
http://www.fr
d try to report.
I don't know if memfd needs to do something special, or just convert!
I really don't know...
> It might also be a good idea to investigate whether one specific meta
> data attachments causes the regression.
Ok will do, thanks!
>
> Thanks,
> Daniel
--
Djalal Harouni
http://opendz.org
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/systemd-devel
On Mon, Jun 30, 2014 at 01:54:57AM +0100, Djalal Harouni wrote:
> On Sun, Jun 29, 2014 at 07:59:38PM -0400, Luke Shumaker wrote:
> > At Sun, 29 Jun 2014 12:31:13 +0100,
> > Djalal Harouni wrote:
> > > On Sat, Jun 28, 2014 at 12:09:56PM -0400, Luke Shumaker wrote:
> &g
On Sun, Jun 29, 2014 at 07:59:38PM -0400, Luke Shumaker wrote:
> At Sun, 29 Jun 2014 12:31:13 +0100,
> Djalal Harouni wrote:
> > On Sat, Jun 28, 2014 at 12:09:56PM -0400, Luke Shumaker wrote:
> > > This is accomplished by having wait_for_container() return a positive
> &
On Fri, Jun 27, 2014 at 01:02:19PM +0200, Daniel Mack wrote:
> On 06/27/2014 12:46 PM, Kay Sievers wrote:
> > On Fri, Jun 27, 2014 at 12:32 PM, Djalal Harouni wrote:
> >> For connections with the KDBUS_HELLO_CACHE_META flag dup the
> >> metadata/credentials from handle
if (r < 0) explicitly set to EXIT_FAILURE, otherwise
return the exit code of the containered process.
> +if (r < 0)
> + r = EXIT_FAILURE;
> break;
> } else if (container_status == CONTAI
On Fri, Jun 27, 2014 at 04:55:30PM -0700, Steven Noonan wrote:
> On Fri, Jun 27, 2014 at 3:14 PM, Djalal Harouni wrote:
> > On Fri, Jun 27, 2014 at 02:28:56PM -0700, Greg KH wrote:
> >> On Fri, Jun 27, 2014 at 10:19:03PM +0100, Djalal Harouni wrote:
> >> > On Fri, Ju
On Fri, Jun 27, 2014 at 02:28:56PM -0700, Greg KH wrote:
> On Fri, Jun 27, 2014 at 10:19:03PM +0100, Djalal Harouni wrote:
> > On Fri, Jun 27, 2014 at 12:23:05PM +0100, Djalal Harouni wrote:
> > > On Fri, Jun 27, 2014 at 01:04:00PM +0200, Daniel Mack wrote:
> > > >
On Fri, Jun 27, 2014 at 12:23:05PM +0100, Djalal Harouni wrote:
> On Fri, Jun 27, 2014 at 01:04:00PM +0200, Daniel Mack wrote:
> > On 06/27/2014 12:51 PM, Djalal Harouni wrote:
> > > Just to let you know that I did notice a regression by ~70% when running
> > > test-kdb
On Fri, Jun 27, 2014 at 01:04:00PM +0200, Daniel Mack wrote:
> On 06/27/2014 12:51 PM, Djalal Harouni wrote:
> > Just to let you know that I did notice a regression by ~70% when running
> > test-kdbus-benchmark on a kvm guest (that's what I've under hands now)
> >
x27;
name_acquire(): flags after call: 0x0
-- entering poll loop ...
stats: 5590 packets processed, latency (usecs) min/max/avg 113/10274/131
stats: 5507 packets processed, latency (usecs) min/max/avg 113/10314/132
stats: 5550 packets processed, latency
Use the new KDBUS_HELLO_CACHE_META flag to improve benchmark by ~50%
This reduces latency and allows sending metadata at the same time.
Signed-off-by: Djalal Harouni
---
test/test-kdbus-benchmark.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/test/test-kdbus
privileged process.
Signed-off-by: Djalal Harouni
---
connection.c | 22 +++---
connection.h | 2 +-
2 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/connection.c b/connection.c
index 85ffa5a..f460ad1 100644
--- a/connection.c
+++ b/connection.c
@@ -1230,7 +1230,16 @@ int
This is a preparation patch thats add kdbus_meta_memdup() to memdup a
metadata object. This is useful to connections with a
KDBUS_HELLO_CACHE_META flag.
Signed-off-by: Djalal Harouni
---
metadata.c | 27 +++
metadata.h | 3 ++-
2 files changed, 29 insertions(+), 1
cause
the source connection to collect its data on every broadcast round.
This allows to provide metadata and reduce latency.
Signed-off-by: Djalal Harouni
---
kdbus.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/kdbus.h b/kdbus.h
index 0b189cb..f637203 100644
--- a/kdbus.h
+++ b
Hi,
First this is just to investigate things! and if it follows kdbus use
cases. This is in the spirit of the late kdbus benchmark threads.
Do not apply, just trying to investigate things and if it is really
useful.
I did benchmarks and the improvement is say ~50%, but did not post
numbers, sorry
On Fri, Jun 27, 2014 at 11:44:06AM +0200, Daniel Mack wrote:
> On 06/27/2014 11:33 AM, Djalal Harouni wrote:
> > On Fri, Jun 27, 2014 at 11:24:48AM +0200, Daniel Mack wrote:
>
> >> No, that's what I'm saying. We cannot cache anything as any information
> >&g
the bus owner nor the bus! you should not connect to
that bus.
I've already patches that do this! will send them in minutes, we have
metadata for free and bechmark improvment hmm say ~50%
>
> Thanks,
> Daniel
>
> ___
Hi,
On Fri, Jun 20, 2014 at 07:12:13PM +0100, Djalal Harouni wrote:
> On Fri, Jun 20, 2014 at 08:01:04PM +0200, Daniel Mack wrote:
> > On 06/20/2014 07:28 PM, Daniel Mack wrote:
> > > On 06/20/2014 06:50 PM, Djalal Harouni wrote:
> > >> Use the db->entries_has
1 - 100 of 262 matches
Mail list logo