Right. this is why it's tricky. You don't want to force an interrupt every
microsecond just because sometimes people might be sleeping with that
granularity (because it would be too often). But the delay is long enough
that a busy-wait is also wrong. You need to dynamically change the clock
behavi
Da Zheng, le Mon 16 Aug 2010 06:02:17 -0400, a écrit :
> Should we still handle the special cases in mach_msg or in another system
> call?
>
> While sleeping in microseconds or nanoseconds, the kernel should do the loop
> on
> the behalf of the user process. Why do we still need a wakeup queue?
olafbuddenha...@gmx.net, le Fri 13 Aug 2010 08:42:33 +0200, a écrit :
> I think improving the timeout granularity in general would be rather
> complicated, and make little sense...
Improving the timeout granularity for even just one case is most of the
work: programming e.g. the HPET timer, managi
Hi,
> I've first seen this problem after having had the following command
> line run for a week, or two, or three:
>
> Start `screen`. Find PID of pfinet.
>
> $ while sleep 66; do echo "$(date)" " $(ps --no-header
> --format=hurd -p [PID])"; done | tee ps-pfinet
>
> Leave it running, deta
Hi,
> `cat < /dev/zero | cat > /dev/null` will eventually make the system
> crash, likewise when using a FIFO. [...]
I think this was the issue where after spending considerable time trying
to isolate it, I finally realized that it actually doesn't happen
anymore with a more recent gnumach... Pro
* Replace $(..)foo with $(top_builddir)/foo for the sake of consistency.
* Define default values for $({top_,}{build,src}dir) and let individual
makefiles handle the special cases (namely, the top Makefile and the ones
which configure generates in the build directory).
* Replace the $(INCLUDES)
Replaces the convoluted attempts at detecting whether config.make was found in
Makeconf by a simple definition in config.make itself.
Signed-off-by: Jeremie Koenig
---
Makeconf |8 ++--
config.make.in |3 +++
2 files changed, 5 insertions(+), 6 deletions(-)
diff --git a/Makeco
Part stores are used for debian-installer, and they're considered more
elegant than kernel partition devices. This patch makes sure they're
included in the build unless the user disables them explicitely.
* configure.in: look for libparted again (see also 7c7a04eb).
* Makeconf: add $(PARTED_LIBS)
* Remove the mostly unused $(doinst).
* Set $(installationdir) directly rather than through $(makemode-instdir).
* Reorganize the logic to be more straightforward.
Signed-off-by: Jeremie Koenig
---
Makeconf | 89 +-
1 files changed, 3
* config.make.in: _set_ the initial value for $({C,CPP,LD}FLAGS).
* Makeconf: alter $(FOOFLAGS) _after_ config.make has been included.
* */Makefile: make sure $(FOOFLAGS) is modified after Makeconf has
been included.
The current situation with {C,CPP,LD}FLAGS is this:
- foo/Makefile may in princ
Hello, me again, these patches are related to the build system.
The first three actually have an effect:
[PATCH 1/7] Fix the build system to handle part stores
[PATCH 2/7] Fix the way FOOFLAGS is handled
[PATCH 3/7] Update ./configure (you may want to do this yourself)
The latter ones are cleanu
The target-specific $($*-CFLAGS) and $($*-CPPFLAGS) are already embedded
into $(CFLAGS) and $(CPPFLAGS), so it's not necessary to add them at the
usage site.
Signed-off-by: Jeremie Koenig
---
Makeconf |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/Makeconf b/Makeconf
Signed-off-by: Jeremie Koenig
---
procfs_pid_files.c | 10 --
1 files changed, 4 insertions(+), 6 deletions(-)
diff --git a/procfs_pid_files.c b/procfs_pid_files.c
index 9dbe3eb..3008145 100644
--- a/procfs_pid_files.c
+++ b/procfs_pid_files.c
@@ -46,14 +46,12 @@ update_pid_entries (s
Process 2 is usually gnumach. This is not perfect, as the format of
/proc/cmdline and /proc/*/cmdline are different on Linux. Namely, the
latter includes NUL bytes to separate subsequent arguments, while the
former contains only spaces and a trailing newline.
Signed-off-by: Jeremie Koenig
---
pr
Signed-off-by: Jeremie Koenig
---
procfs_pid_files.c | 11 ++-
1 files changed, 10 insertions(+), 1 deletions(-)
diff --git a/procfs_pid_files.c b/procfs_pid_files.c
index 3008145..e2ad2fd 100644
--- a/procfs_pid_files.c
+++ b/procfs_pid_files.c
@@ -49,7 +49,16 @@ update_pid_entries (
Among other things, many parts of Debian-installer want /proc/mounts.
Implementing it the right way is non-trivial, though. In the meantime,
we publish a symlink to /etc/mtab so that the user can fake it in
whatever way they want. Furthermore I hope to be porting busybox mount
soon, and use it to m
On Linux, /proc//cmdline is a NUL-separated list of arguments.
We used to truncate after the first one and add some whitespace.
Signed-off-by: Jeremie Koenig
---
procfs_pid_files.c | 14 +++---
1 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/procfs_pid_files.c b/proc
Hi, here comes the latest (hopefully final) version of my procfs series.
Changes:
* fix the typo;
* split the last patch in two.
Could someone consider you merge this?
Hi,
Should we still handle the special cases in mach_msg or in another system call?
While sleeping in microseconds or nanoseconds, the kernel should do the loop on
the behalf of the user process. Why do we still need a wakeup queue?
Best,
Zheng Da
On 8/13/10 5:28 PM, Thomas Bushnell, BSG wrote:
19 matches
Mail list logo
