On Sun, Aug 2, 2020 at 8:32 AM Waldek Kozaczuk wrote:
> This is possibly one of the longest emails I have ever written. So bear
> with me ...
>
I'll try :-) Replies inline:
> After all the recent patches, I have submitted that replaced many source
> files under libc/ (~70) and headers under in
It seems you already did this change in a bigger patch and this one should
be dropped, right?
--
Nadav Har'El
n...@scylladb.com
On Sat, Aug 1, 2020 at 8:09 AM Waldemar Kozaczuk
wrote:
> Signed-off-by: Waldemar Kozaczuk
> ---
> include/api/
iases.c) allows us to
drop a change from musl (here res_init.c).
Signed-off-by: Nadav Har'El
---
Makefile| 3 ++-
libc/aliases.c | 31 +++
libc/network/res_init.c | 7 ---
3 files changed, 33 insertions(+), 8 deletions(-)
create mode 1
I want to later create an alternative patch to this that uses a linker
script (e.g., libc/aliases.ld) to copy symbols - and then we can decide
which approach is nicer (or less ugly).
--
Nadav Har'El
n...@scylladb.com
On Tue, Aug 4, 2020 at 1:40 AM Nadav Har'El wrote:
> For a long
es, wrappers in libc/math/aliases.cc)
to the new aliases.ld.
Signed-off-by: Nadav Har'El
---
Makefile| 2 +-
arch/aarch64/loader.ld | 1 +
arch/x64/loader.ld | 1 +
libc/aliases.ld | 12
libc/network/res_init.c | 7 ---
5 files changed, 15 i
On Tue, Aug 4, 2020 at 1:41 AM Nadav Har'El wrote:
> I want to later create an alternative patch to this that uses a linker
> script (e.g., libc/aliases.ld) to copy symbols - and then we can decide
> which approach is nicer (or less ugly).
>
>
Done. Just sent a v2 with a l
On Wed, Aug 5, 2020 at 7:36 AM Waldek Kozaczuk wrote:
> This looks like a great solution. I was about to commit but I realized
> that alias symbols like __res_init become global symbols vs weak in
> loader.elf. Does it have any practical consequences to how they are handled
> by OSv dynamic link
On Thu, Aug 6, 2020 at 1:00 AM Rick Payne wrote:
>
> I've been noodling with the old 'assigned virtio' code, trying to make
> it work again so I can use this method to get raw packets. See
> discusion on the 'raw sockets' thread.
>
> I'm mostly there (though I will admit, I'm far from a C++ progr
On Thu, Aug 6, 2020 at 11:02 AM Nadav Har'El wrote:
> The assert was there to make a rare random crash into a reliable crash.
> The thing is
> that symbol resolution needs to take a mutex, because in very rare cases
> it might be
> running in parallel to someone loading
On Mon, Aug 10, 2020 at 8:02 AM Waldemar Kozaczuk
wrote:
> In more complicated situations when single musl file
> calls syscall() for 2 or more system call types (like stdio/tmpnam.o),
> we cannot have single macro that will solve it like previous patch
> did.
>
> In those cases we fallback to st
On Mon, Aug 10, 2020 at 5:17 PM Waldek Kozaczuk
wrote:
> Any reason why we use __syscall(SYS_close, ...) vs close()?
>
> struct if_nameindex *if_nameindex()
> {
> size_t n;
> void *p = 0;
> int s = socket(AF_INET, SOCK_DGRAM|SOCK_CLOEXEC, 0);
> if (s>=0) {
> for (n=0; (p=do_nameindex(s, n)) == (v
On Mon, Aug 10, 2020 at 6:50 PM Waldek Kozaczuk
wrote:
>
> There is one thing to discuss. While I agree that we should use '
> --include libc/*syscall_to*_function.h' for all musl source and have
> compiler detect any undefined individual translation macros (like
> __OSV_TO_FUNCTION_SYS_clock_get
On Mon, Aug 10, 2020 at 5:21 PM Waldek Kozaczuk
wrote:
> Can anyone explain the motivation behind this script and how the generated
> alltypes.h is intended to be used?
I can try. I think we took this idea from Musl.
The C library has type definitions in a lot of different header files. For
ex
Don't you also need to add the alias __duplocale?
On Sat, Aug 15, 2020, 07:27 Waldemar Kozaczuk wrote:
> Signed-off-by: Waldemar Kozaczuk
> ---
> Makefile| 4 ++--
> libc/aliases.ld | 1 +
> libc/locale/duplocale.c | 15 ---
> libc/locale/setlocale.c | 13
On Thu, Aug 20, 2020 at 7:18 AM Commit Bot wrote:
> From: Waldemar Kozaczuk
> Committer: Waldemar Kozaczuk
> Branch: master
>
> tests: adapted w/ctype tests to work with boost unittest and added
> tst-string
>
> Signed-off-by: Waldemar Kozaczuk
>
> ---
> diff --git a/modules/tests/Makefile b/m
// looks like a large number, but gets treated as
"-1".
uiomove(..., x)// x sign extended to 64-bit "-1", i.e.,
18446744073709551616, will refuse to work because > SSIZE_MAX, so good.
--
Nadav Har'El
n...@scylladb.com
On Thu, Aug 20, 2020 at 7:18 AM Commit B
I will commit, but I think you forgot to patch the makefile to actually
build this test? If that's true please send a followup patch.
--
Nadav Har'El
n...@scylladb.com
On Wed, Aug 26, 2020 at 6:54 AM Waldemar Kozaczuk
wrote:
> This patch adds new unit test tst-time.cc that te
On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Initialize Xen event interrupt separately from xenbus driver, so that
> events
> can be served before xenbus is up, or even in case when xenbus not being
> probed
> at all or not required, e.g
On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Xen platform PCI is x86-specific feature, not currently present on aarch64.
>
> For xenbus to work on both platforms the generic xenbus driver should be
> probed
> directly (under aarch64 in c
Reviewed-by: Nadav Har'El
On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Features used by various Xen drivers, i.e. xenbus.
>
> Signed-off-by: Sergiy Kibrik
> ---
> arch/aarch64/cpuid.cc | 1
On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Complete driver to properly work with xl console utility, i.e. full
> read/write
> access. Driver is extended to work with shared structure (ring) and event
> channel.
>
> Signed-off-by: Sergi
Just a few lines above, we had exactly the same code... No need to do
it again.
Signed-off-by: Nadav Har'El
---
libc/signal.cc | 4
1 file changed, 4 deletions(-)
diff --git a/libc/signal.cc b/libc/signal.cc
index b7351ee..832002c 100644
--- a/libc/signal.cc
+++ b/libc/sign
Stop using OSv-specific features (namely debug()) in this test, so we
can also run it on Linux.
This test currently hangs on Linux, because the test assumes that signal()
sets system calls to not restart - while on Linux, they do.
Signed-off-by: Nadav Har'El
---
tests/tst-kill.cc
-off-by: Nadav Har'El
---
tests/tst-kill.cc | 12 ++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/tests/tst-kill.cc b/tests/tst-kill.cc
index fa99eb8..3d4f942 100644
--- a/tests/tst-kill.cc
+++ b/tests/tst-kill.cc
@@ -78,8 +78,16 @@ int main(int ac, char** av)
r
We tested that killing a non-existant process 17 resulted in an error.
But on Linux, it is quite possible that process 17 does exist, and this
test will fail. So change it to an even less likely to exist number.
Signed-off-by: Nadav Har'El
---
tests/tst-kill.cc | 2 +-
1 file chang
Looks good. I wonder how I tested the original code ;-)
I think my original intention was to do an lseek() - once - instead of
O_APPEND - but this lseek is missing. O_APPEND should work too, it's
somewhat of an overkill I guess, but should work. I'll commit this.
--
Nada
On Mon, Mar 6, 2017 at 4:08 PM, Sergiy Kibrik wrote:
> On 03/05/2017 11:08 AM, Nadav Har'El wrote:
>
>
> On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
> osv-dev@googlegroups.com> wrote:
>
>> Initialize Xen event interrupt sep
On Mon, Mar 6, 2017 at 5:24 PM, Sergiy Kibrik wrote:
>
> On 03/05/2017 12:45 PM, Nadav Har'El wrote:
>
>
> +
>
>> void XEN_Console::write(const char *str, size_t len) {
>> - HYPERVISOR_console_write(str, len);
>> +
On Mon, Mar 13, 2017 at 12:38 PM, Sergiy Kibrik <
sergiy.kib...@globallogic.com> wrote:
> On 03/06/2017 05:24 PM, Sergiy Kibrik wrote:
>
> +return _interface->in_cons != _interface->in_prod;
>> }
>>
>> char XEN_Console::readch() {
>> -return '\0'; /*TODO: support input */
>> +XENCONS
: Nadav Har'El
---
drivers/line-discipline.cc | 52 --
1 file changed, 27 insertions(+), 25 deletions(-)
diff --git a/drivers/line-discipline.cc b/drivers/line-discipline.cc
index 4e21769..082c392 100644
--- a/drivers/line-discipline.cc
+++ b/drivers
On Mon, Mar 13, 2017 at 4:21 PM, Justin Cinkelj
wrote:
> Based on what open-mpi does.
>
Did you copy code from open-mpi? Is the license compatible? Should we add a
copyright comment?
Also, what I see below doesn't seem to be a test - how can it fail?
Maybe you want it in tests/misc-something, i
On Mon, Mar 13, 2017 at 4:32 PM, Nadav Har'El wrote:
>
>
> We need to understand how IFF_SLAVE and IFF_SIMPLEX came to have the same
> bit... Did we get one from BSD and one from Linux?
>
Indeed this seems to be the case. The bits which Linux applications expect
are in /usr
me bit position, 0x800). In particular OpenMPI did not like
seeing this flag, and refused to use this network interface.
Fixes #862
Signed-off-by: Nadav Har'El
---
bsd/sys/compat/linux/linux_ioctl.cc | 19 ++-
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/
OSv), the new names do not work and we need to use the old ones
from before commit adf78de4 - otherwise we get errors on "connect".
So in this patch we try both names, trying the new names first, and if
they fail, the old names.
Signed-off-by: Nadav Har'El
---
scripts/loader.py |
e.
Signed-off-by: Nadav Har'El
---
scripts/loader.py | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/scripts/loader.py b/scripts/loader.py
index b16538c..3a52377 100644
--- a/scripts/loader.py
+++ b/scripts/loader.py
@@ -245,7 +245,10 @@ def vma_list(node=None)
On Thu, Mar 16, 2017 at 2:15 PM, wrote:
> Hi,
>I have two questions.
>
> First, what are the features and components present in cloudius/osv-base
> base image?
>
This and other "Capstan base images" are built by the script
scripts/build-capstan-images.
As you can see there, it includes the
On Fri, Dec 23, 2016 at 6:44 PM, Rean Griffith
wrote:
> Fixed and resubmitted.
>
> Thanks,
> Rean
>
Hi, we see sporadic hangs of the pthread_barrier test in - I opeened an
issue: https://github.com/cloudius-systems/osv/issues/857
Because I cannot reproduce this myself, I don't know yet if this
On Mon, Mar 20, 2017 at 4:12 PM, Sergiy Kibrik <
sergiy.kib...@globallogic.com> wrote:
> On 03/20/2017 03:55 PM, Sergiy Kibrik wrote:
>
>> void XEN_Console::write(const char *str, size_t len) {
>> - HYPERVISOR_console_write(str, len);
>> +assert(len > 0);
>> +if (!_interface) {
>> +
On Mon, Mar 20, 2017 at 6:38 PM, Sergiy Kibrik <
sergiy.kib...@globallogic.com> wrote:
>
> +XENCONS_RING_IDX prod = _interface->out_prod;
>
> +constexpr auto ringsize = sizeof(_interface->out) - 1;
>>> +while (true) {
>>> +XENCONS_RING_IDX cons = _interface->out_cons;
>>>
>> +
Hi,
I cannot get this patch to apply cleanly after applying patches 1-3. Can
you please send a new version of this patch? Thanks.
--
Nadav Har'El
n...@scylladb.com
On Fri, Mar 3, 2017 at 11:45 AM, 'Sergiy Kibrik' via OSv Development <
osv-dev@googlegroups.com> wrote:
>
On Tue, Mar 21, 2017 at 4:05 PM, wrote:
> I am mailing about OSv Redis.
>
> I have noticed in the download section http://osv.io/downloads/ that ova
> for Redis contains quite old 0.24 version.
>
All our prepared images are old. I'm CC'ing Peka to consider if he wants to
change that.
You can bu
On Thu, Feb 23, 2017 at 3:54 PM, Waldemar Kozaczuk
wrote:
> New ec2-make-ami.py uses boto3 to interact with AWS. Also it logs time it
> takes to execute
> all steps involved in creating an AMI. It also accepts new parameter -i
> that allows
> to specify location of the OSv image file.
>
Hi, I kn
On Tue, Mar 21, 2017 at 5:50 PM, Nadav Har'El wrote:
>
>
> On Tue, Mar 21, 2017 at 4:05 PM, wrote:
>
>> I am mailing about OSv Redis.
>>
>> I have noticed in the download section http://osv.io/downloads/ that ova
>> for Redis contains quite old 0.24 ve
On Wed, Mar 22, 2017 at 11:47 AM, Nadav Har'El wrote:
>
>
> On Tue, Mar 21, 2017 at 5:50 PM, Nadav Har'El wrote:
> While redis 3.0.1 still seems to work correctly, redis 3.2.8 (which you
> can get by trivially changing the version in apps/redis-memonly/GET) does
> n
Thanks. Looks good to me. Justin, can you please also have a look?
--
Nadav Har'El
n...@scylladb.com
On Wed, Mar 22, 2017 at 2:57 PM, Waldemar Kozaczuk
wrote:
> From: wkozaczuk
>
> The patch addressing issue #816 changed dhcp_mbuf::compose_request()
> method to
> make it
Thanks. I'm really not versed enough in the DHCP RFC to make much sense out
of this. Waldemar, can you please review? Thanks.
--
Nadav Har'El
n...@scylladb.com
On Thu, Mar 23, 2017 at 12:41 AM, Justin Cinkelj
wrote:
> The patch works OK for me, tested agains dnsmasq.
>
On Wed, Mar 22, 2017 at 7:30 PM, wrote:
> Hi Nadav,
>
> Thanks for your effort and such a quick response. Much appreciate it.
>
> I am more than happy to give that a go but will need a short guide from
> you. Basically I have only downloaded your virtual appliance of redis and
> it has worked as
On Fri, Mar 24, 2017 at 10:37 PM, Justin Cinkelj
wrote:
> If we agree I correctly interpreted the part of RFC in previous mail, that
> patch should be included too?
> BR Justin
>
Yes, thanks. I committed this too.
--
You received this message because you are subscribed to the Google Groups "OS
es between the different
versions' patches should disappear. Rean, can you please fix that and also
see my other comments, and submit a new patch? Thanks.
>
> On Monday, February 6, 2017 at 4:13:45 AM UTC-5, Nadav Har'El wrote:
>>
>>
>> On Mon, Feb 6, 2017 at 10:27 AM
tions
by a little bit, but hopefully they are not heavily used by performance-
minded applications, anyway.
Fixes #867.
Signed-off-by: Nadav Har'El
---
Makefile | 14 ++
1 file changed, 14 insertions(+)
diff --git a/Makefile b/Makefile
index 5133d8e..ba84338 100644
--- a/M
Were you able to reproduce this problem - perhaps by reducing sleeping in
the test case, or something?
Some comments below.
On Mon, Apr 3, 2017 at 4:30 AM, 'rean' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Signed-off-by: Rean Griffith
> ---
> include/osv/latch.hh| 3 +++
> l
If I understand correctly, the patch from Waldemar Kozaczuk which I just
applied makes this one unnecessary, and I shouldn't try to apply this.
Right?
--
Nadav Har'El
n...@scylladb.com
On Thu, Mar 30, 2017 at 4:29 PM, Justin Cinkelj
wrote:
> The old hostname (DNS name)
On Wed, Mar 29, 2017 at 3:54 PM, Justin Cinkelj
wrote:
> When command is run via CLI (run.py -e "CMD"), the CMD string can contain
> $KEY, and KEY will be replaced with corresponding value from environ.
> The patch implements this also for commands run via REST or via runscript.
>
> Fixes #868
>
orrect mess, I agree that it's better to fix it even
of the cost of increasing the mess - and we can clear the mess later.
--
Nadav Har'El
n...@scylladb.com
On Thu, Apr 6, 2017 at 8:42 AM, 'rean' via OSv Development <
osv-dev@googlegroups.com> wrote:
> Signed-o
On Thu, Apr 6, 2017 at 10:58 AM, Nadav Har'El wrote:
>
> I also have a general worry that this code has become much more complex
> than it really needs to be - for example do we really need three separate
> mutexes (two in this code, one inside the latch) to implement a bar
e only a single
mutex, and a single condition variable (actually, a waitqueue which is
like a condition variable but does not require an additional internal
mutex in its implementation).
Signed-off-by: Nadav Har'El
---
libc/pthread_barrier.cc | 117 +---
On Sun, Apr 9, 2017 at 5:33 PM, Nadav Har'El wrote:
> A third alternative to re-implementing from scratch or fixing this
> latch-based implementation, could be to implement based on other building
> blocks that OSv already has (and are better debugged than either "l
On Tue, Apr 11, 2017 at 8:13 AM, Rean Griffith
wrote:
> Hi Nadav,
>
> I looked at your patch and it looks good.
>
> I tried running the tests I have been using locally to trigger hangs and
> it passed - no hangs observed thus far after running a few hundred
> repeated rounds with between 1 and 10
On Wed, Apr 12, 2017 at 11:09 AM, Nadav Har'El wrote:
> On Tue, Apr 11, 2017 at 8:13 AM, Rean Griffith
> wrote:
>
>> Hi Nadav,
>>
>> I looked at your patch and it looks good.
>>
>> I tried running the tests I have been using locally to trigger hangs a
On Mon, Apr 10, 2017 at 1:05 AM, Rick Payne (Offshore) wrote:
> Hi,
>
> Is there a good example that uses the virtio network interfaces direct?
> ie. an application that uses ‘assign-net’ and talks directly to the network
> interface code in OSv?
>
> I’m looking for a sample I can learn from.
>
On Wed, Apr 12, 2017 at 8:36 PM, Dor Laor wrote:
> We did hack memcache to get packets directly from the device. You can
> search for this example, it might be good enough for you
>
>
What Dor is referring to is https://github.com/vladzcloudius/osv-memcached
(you can build it with "scripts/build
ost 0.3 seconds for 32 vcpus. However, this "stuck time" does not cause
any real problems as code which needs functional clocks and preemptive
multitasking will run later anyway.
Signed-off-by: Nadav Har'El
---
drivers/clock-common.hh | 2 +-
drivers/clock-common.cc | 10
On Fri, Apr 14, 2017 at 3:12 PM, Waldek Kozaczuk
wrote:
> Hey,
>
> Can I get access to the blog github or wherever its content is stored so I
> can fix it?
>
> Thanks,
> Waldek
>
>
Thanks to Pekka and Tzach, we finally found the "missing" blog repository
and fixed the layout issue:
It turns out
On Sun, Apr 23, 2017 at 2:06 PM, Avi Kivity wrote:
> It could be the new ABI, see https://gcc.gnu.org/
> onlinedocs/libstdc++/manual/using_dual_abi.html.
>
>
> Try compiling your code with _GLIBCXX_USE_CXX11_ABI=0.
>
Indeed. Check out also commit 6a3bff38a281e65ee715bab4fadef63e0918f7d3
where we
On Sun, Apr 23, 2017 at 4:23 PM, Rick Payne (Offshore) wrote:
>
> > On 23 Apr 2017, at 05:14, Nadav Har'El wrote:
> >
> > Pekka, another question we should ask ourselves is what is the plan with
> Capstan, and more importantly, the binary OSv releases. The OSv v
Thanks. I opened https://github.com/cloudius-systems/osv/issues/871 to
track this issue.
--
Nadav Har'El
n...@scylladb.com
On Tue, Apr 25, 2017 at 12:42 PM, Boris Brankovic <
brankovic.boris...@gmail.com> wrote:
> Hi all,
>
> I am trying to test a Java Spring-Boot Applicati
Remove some dead code (an ifdef'ed out static function which was never used)
from fs/vfs/main.cc.
Signed-off-by: Nadav Har'El
---
fs/vfs/main.cc | 15 ---
1 file changed, 15 deletions(-)
diff --git a/fs/vfs/main.cc b/fs/vfs/main.cc
index 924bfad..44ae642 100644
--- a/fs/v
mfs
Fixes #871
Signed-off-by: Nadav Har'El
---
scripts/mkbootfs.py | 20 +++-
fs/ramfs/ramfs_vnops.cc | 44
fs/vfs/main.cc | 26 --
usr_nozfs.manifest.skel | 3 +--
4 files changed, 80 in
image it may not work because of a silly
limit we have in place on the size of the kernel (which includes the ram
disk) - the error message will tell you how to fix it.
Can you please give this patch a try and see how it works out for you?
Thanks,
Nadav.
--
Nadav Har'El
n...@scylladb.co
On Wed, Apr 26, 2017 at 11:53 AM, Benoît Canet wrote:
>
>
> 2017-04-26 0:50 GMT+02:00 Nadav Har'El :
>
ramfs_rmdir(struct vnode *dvp, struct vnode *vp, char *name)
>
>> @@ -388,11 +424,11 @@ ramfs_rename(struct vnode *dvp1, struct vnode *vp1,
>>
On Thu, Apr 27, 2017 at 7:37 AM, Waldek Kozaczuk
wrote:
>
> I am also thinking of changing my original plan. Instead of porting all
> drivers first I will first port vmbus only and then work on implementing
> hyperv::vmbus
> and hyperv::hyperv_driver classes to make OSv at least boot on Hyper/V
On Thu, Apr 27, 2017 at 11:37 AM, Avi Kivity wrote:
>
>
> On 04/27/2017 07:37 AM, Waldek Kozaczuk wrote:
>
> Here are other issues where I am seeking advice:
>
>- *SYSCTL* in FreeBSD provides configuration management mechanism for
>kernel and device drivers (https://www.freebsd.org/cgi/
>
On Thu, Apr 27, 2017 at 10:46 AM, Boris Brankovic <
brankovic.boris...@gmail.com> wrote:
>
> Thank you very much for fixing this issue! The build command works fine
> now.
>
> I just get an error if I want to start the unikernel with the
> "run.py"-script.
>
> The error says:
> OSv v0.24-327-gd6ca
On Thu, Apr 27, 2017 at 12:18 PM, Nadav Har'El wrote:
>
> On Thu, Apr 27, 2017 at 10:46 AM, Boris Brankovic <
> brankovic.boris...@gmail.com> wrote:
>
>>
>> Thank you very much for fixing this issue! The build command works fine
>> now.
>>
>&g
On Sun, May 7, 2017 at 9:09 AM, Waldek Kozaczuk
wrote:
> Hyper/V VMBUS seems be using inter processor interrupts delivered by
> *Synthetic
> Interrupt Controller (SynIC). *
>
> The BSD drivers use lapic_ipi_alloc() which per this patch description (
> https://reviews.freebsd.org/D2042) dynamical
might not be
necessary after I added support for nearbyint(). I also had several other
reservations on that patch, which is why it wasn't committed.
Did you also try more recent versions and failed? Or did you really want to
run 6.10 in particular?
--
Nadav Har'El
n...@scylladb.com
On Mon, May 8, 2017 at 2:48 PM, Nadav Har'El wrote:
> Thanks. I committed this.
>
> In February, Rean Griffith sent patch titled
> "Generalize
> nodejs build to support recent versions (6.9.5 LTS and 7.5.0 latest)" which
> enabled even later versions. He ended
On Wed, May 10, 2017 at 1:54 AM, Waldek Kozaczuk
wrote:
> I also hoped it would be possible to boot simple "hello world" image on
> Hyper/V without adding anything to OSv to support Hyper/V drivers. But
> based on the experiments I conducted yesterday it looks like I will have to
> add some minim
On Wed, May 10, 2017 at 8:14 AM, Nadav Har'El wrote:
>
>
> On one hand, tha't nice progress - you can see actual OSv code getting
> printed on hyper-v :-)
> On the other hand, sad that HPET is not supported as a fallback. I looked
> at what Linux did when hyperv fir
On Fri, May 12, 2017 at 10:26 PM, Waldek Kozaczuk
wrote:
> Given I am almost done porting the FreeBSD drivers code for Hyper/V (at
> least the VMBus part) I have decided to proceed with the implementation of
> the clock using Hyper/V time synchronization service. This will obviously
> require put
=ramfs
(note the fs=ramfs, causing all the files to be included in the kernel
executable)
fails with
strip: 'libosv.so': No such file
This patch simply removes libosv.so from the java module. Nothing seems
to rely on it.
Signed-off-by: Nadav Har'El
---
modules/java-base/u
er reading a certain length),
and in Makefile (which verifies that the kernel isn't too large).
Fixes #872.
Signed-off-by: Nadav Har'El
---
fastlz/lzloader.cc | 10 --
Makefile| 16 +---
arch/x64/boot16.S | 2 +-
arch/x64/lzloa
Hi,
On Fri, May 19, 2017 at 7:52 AM, Waldek Kozaczuk
wrote:
> I finally managed to fully boot and execute simple C hello work program on
> Hyper/V.
>
That's great!
With the patch I sent yesterday allowing a kernel of any size, you should
now be able to run any OSv image when built with fs=ramfs
this as a "TODO" and only go back to it after you make the network (for
example) work.
--
Nadav Har'El
n...@scylladb.com
On Fri, May 19, 2017 at 6:03 PM, Waldek Kozaczuk
wrote:
> The essence of my changes is this new clock class - https://github.com/
> wkozaczuk/osv/bl
On Fri, May 19, 2017 at 11:54 PM, Waldek Kozaczuk
wrote:
> What do you mean by "exit"? Is it because it calls a function? Would
> inlining help?
>
"VMX" (virtual machine extensions) is Intel's hardware support for
virtualization. What makes it efficient is that the guest code mostly runs
normall
Hi, sorry for the belated review.
There's a lot of stuff here. Do you need all of that just for the "hello
world" example, or are you adding much of it in anticipation of the
network/disk support later?
On Fri, May 26, 2017 at 8:37 AM, Waldemar Kozaczuk
wrote:
> - added subset of original Free
Looks mostly good. Some minor comments below.
On Fri, May 26, 2017 at 8:37 AM, Waldemar Kozaczuk
wrote:
> Here is the list of specific changes:
> - where required by C++ compiler applied appropriate C++ cast
> (static_cast, reinterpret_cast, const_cast)
> - where necessary replaced FreeBSD he
Looks great, thanks.
Some minor comments below.
On Fri, May 26, 2017 at 8:37 AM, Waldemar Kozaczuk
wrote:
> These changes are enough to allow execution of simple OSv apps that do NOT
> need
> networking nor storage access.
Excellent!
> The image built using the command below should
> success
On Thu, Jun 8, 2017 at 7:13 AM, Justin Cinkelj
wrote:
> Gasper in CC original console output:
> OSv v0.24-364-g3cb3ccc
> eth0: 10.32.3.20
> java.so: Starting JVM app using: io/osv/isolated/RunIsolatedJvmApp
> java.so: Setting Java system classloader to IsolatingOsvSystemClassLoader
> and logging
returned fname and base to the object's values when the address
belongs to this object - even if not to any known symbol, because we need
this behavior in dladdr().
Fixes #878, but openjdk8-fedora+tomcat still does not work, because
of yet another bug.
Signed-off-by: Nadav Har'El
---
core/el
locks once for
the entire readv() scattered buffer, as expected.
So we don't need this workaround, which can only hurt performance, and in
the present case, prevent fread() from working on a block device.
Fixes #877
Signed-off-by: Nadav Har'El
---
tests/tst-fread
On Tue, Jun 13, 2017 at 5:54 PM, Justin Cinkelj
wrote:
> Patch seems to work, please apply.
>
> Instead of assert(), libcdio now complains if I ask to extract invalid
> filename.
>
If you tried to read a non-existant file, it should have also failed this
way previously, and not with an assert.
On Wed, Jun 14, 2017 at 4:56 PM, Justin Cinkelj
wrote:
> The calloc(nmemb, size=0) failed due to division by 0 in
> "std::numeric_limits::max() / size".
> Fix this by checking for 0 value before division.
>
> Included is also a small test case for calloc and memset.
>
> Signed-off-by: Justin Cink
Hi, sorry about the delayed response...
On Wed, May 31, 2017 at 6:52 AM, Waldek Kozaczuk
wrote:
>
> bool
>
>> -hyperv_is_timecount_available() {
>>> +hyperv_is_timecount_available(void) {
>>> return (hyperv_features & CPUID_HV_MSR_TIME_REFCNT);
>>>
>>
>> This function needs to run after "h
in my last response which I would
>>> like you to answer before I can send new updated patch.
>>>
>>> On Tuesday, May 30, 2017 at 11:52:32 PM UTC-4, Waldek Kozaczuk wrote:
>>>>
>>>> Please see my answers below.
>>>>
>>>> On Mon,
Hi, sorry about the delay. Hopefully I found the right questins now:
On Wed, May 31, 2017 at 6:29 AM, Waldek Kozaczuk
wrote:
> Please see my answers below. Overall I correct what needed and I will send
> the updated patch.
>
>
>
>>> -static bool
>>> +bool
>>> hyperv_identify(void)
>>> {
>>>
On Thu, Jun 22, 2017 at 7:51 AM, Waldek Kozaczuk
wrote:
> My objective is to push a growing application log file from OSv instance
> into ELK stack (elasticsearch, logstash, kibana) or Splunk. And I am
> looking for best way to do it.
>
> Here are the alternatives:
>
>1. Externally (outside o
k" in the reciever). When TCP_CORK is used and
released after the second write, it gets sent immediately.
We saw these 100ms "hangs" of TCP sessions in real benchmarks of nginx,
because its usage of TCP_CORK was silently ignored.
Fixes #883.
Signed-off-by: Nadav Har'El
---
bsd/sys/
same test for TCP_CORK, it's just that with TCP_NODELAY we
don't need to "uncork" - each small write sends its data immediately so
right after the second write, the reader can almost immediately read all
the data).
Signed-off-by: Nadav Har'El
---
bsd/sys/compat/l
On Fri, Jun 30, 2017 at 10:43 PM, Waldek Kozaczuk
wrote:
> I am not sure if this is the best forum as it relates to the MikelAngelo
> capstan.
>
I think it is, because all the relevant Mikelangelo developers are on this
list, and anyway the "original" Capstan is more-or-less dead, and it's good
301 - 400 of 1373 matches
Mail list logo