> By the way, can you please try to insteaapp of the
>
> assert(sched::preemptable());
>
> try to
>if (!sched::preemptable()) {
> debug_always("preempt counter is %d\n", sched::get_preempt_counter());
>}
>
> Is this counter something sensible, like 1, or some random crap indicat
On Fri, Nov 25, 2016 at 6:13 PM, Justin Cinkelj
wrote:
> Fixes #761
>
> Signed-off-by: Justin Cinkelj
> ---
> libc/pthread.cc | 64 ++
> +--
> 1 file changed, 39 insertions(+), 25 deletions(-)
>
> diff --git a/libc/pthread.cc b/libc/pthread.cc
From: Nadav Har'El
Committer: Nadav Har'El
Branch: master
update apps/ submodule
Signed-off-by: Nadav Har'El
---
diff --git a/apps b/apps
--- a/apps
+++ b/apps
@@ -1 +1 @@
-Subproject commit f8731d0253a7537a15c12321151ef8c6168ad902
+Subproject commit 8b8bf4d1a39ced0c6f7c4433723e3262e0d28a04
From: Rick Payne
Committer: Nadav Har'El
Branch: master
Ensure that the argument to osv_system() is saved in allocated memory
before we spawn the pthread
Signed-off-by: Rick Payne
Message-Id: <1480355794-10065-1-git-send-email-ri...@rossfell.co.uk>
---
diff --git a/erlang/epmd.patch b/erl
From: Rick Payne
Committer: Nadav Har'El
Branch: master
Ensure erlang builds referencing the openssl install that is copied into
the resulting image
Signed-off-by: Rick Payne
Message-Id: <1480355992-10295-1-git-send-email-ri...@rossfell.co.uk>
---
diff --git a/erlang/Makefile b/erlang/Mak
On Mon, Nov 28, 2016 at 6:07 PM, Rick Payne wrote:
>
> > I don't think it should, after we fixed - long ago - the resolution
> order in commit 82513d413502457c8280c6582876653527b9dd1d.
> > Basically, OSv is searched last; If the application defines, say, the
> wait_until() function and uses it, t
The three symbols in elveldb.so matching the prefix you mentioned are (from
nm -D eleveldb.so):
0004e510 W
_ZNSt8_Rb_treeIPN7leveldb12FileMetaDataES2_St9_IdentityIS2_ENS0_10VersionSet7Builder13BySmallestKeyESaIS2_EE8_M_eraseEPSt13_Rb_tree_nodeIS2_E
0005e730 W
_ZNSt8_Rb_treeIPN7leve
On Mon, Nov 28, 2016 at 10:37 PM, Rick Payne wrote:
>
> > On 28 Nov 2016, at 15:52, Rick Payne wrote:
> >
> > eleveldb.so:
> >
> > 002a2560 036b0007 R_X86_64_JUMP_SLO 000617b0
> _ZN7leveldb11GetVarint + 0
> > 002a2568 00340007 R_X86_64_JUMP_SLO
> _ZNSt7
> On 28 Nov 2016, at 15:52, Rick Payne wrote:
>
> eleveldb.so:
>
> 002a2560 036b0007 R_X86_64_JUMP_SLO 000617b0
> _ZN7leveldb11GetVarint + 0
> 002a2568 00340007 R_X86_64_JUMP_SLO
> _ZNSt7__cxx1112basic_s@GLIBCXX_3.4.21 + 0
> 002a2570 022c000
From: Rick Payne
Committer: Nadav Har'El
Branch: master
shmget: Handle calls for non-existant object without IPC_CREAT
Signed-off-by: Rick Payne
Message-Id:
---
diff --git a/libc/shm.cc b/libc/shm.cc
--- a/libc/shm.cc
+++ b/libc/shm.cc
@@ -99,6 +99,8 @@ int shmget(key_t key, size_t size, in
> On 28 Nov 2016, at 10:00, Avi Kivity wrote:
>
> Can you provide the output of "readelf --relocs" for your NIF .so, section
> .rela.plt, around the 171st entry? Unfortunately I can't tell you which NIF
> triggered it (or if the output will be useful), but it's worth trying.
There are only 2
Signed-off-by: Rick Payne
---
erlang/Makefile | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/erlang/Makefile b/erlang/Makefile
index 3e2bfa4..324b05f 100644
--- a/erlang/Makefile
+++ b/erlang/Makefile
@@ -17,7 +17,7 @@ otp_src_$(OTP_VERSION): otp_src_$(OTP_VERSION).tar.gz
Signed-off-by: Rick Payne
---
erlang/epmd.patch | 21 -
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/erlang/epmd.patch b/erlang/epmd.patch
index 7e4ac61..dc9e3b6 100644
--- a/erlang/epmd.patch
+++ b/erlang/epmd.patch
@@ -1,6 +1,5 @@
-diff -up erts/etc/common/
Hi,
> Thanks. By the way, another difference between this osv_system() and the
> usual system() is that system() waits for the command to finish, while
> osv_system() apparently does not. Does this not bother anything?
I don’t think so.
> Thanks. As I requested in another mail, can you please
> I don't think it should, after we fixed - long ago - the resolution order in
> commit 82513d413502457c8280c6582876653527b9dd1d.
> Basically, OSv is searched last; If the application defines, say, the
> wait_until() function and uses it, this definition will be used and mask
> OSv’s.
Ok.
> I
On Sun, Nov 27, 2016 at 11:02 PM, Rick Payne wrote:
> The patch to erlexec in apps/erlang/epmd.patch can have an issue where the
> argument to osv_sh() can have disappeared from the callee’s stack by the
> time the thread is created. The simple solution is to strsave() it in
> osv_system.
>
Than
Signed-off-by: Rick Payne
---
libc/shm.cc | 2 ++
1 file changed, 2 insertions(+)
diff --git a/libc/shm.cc b/libc/shm.cc
index 7c1a11a..d5647df 100644
--- a/libc/shm.cc
+++ b/libc/shm.cc
@@ -99,6 +99,8 @@ int shmget(key_t key, size_t size, int shmflg)
fdesc f(fref);
On Mon, Nov 28, 2016 at 5:29 PM, Rick Payne wrote:
>
> > On 28 Nov 2016, at 10:57, Nadav Har'El wrote:
> >
> > However, our implementation of the symbol resolution can sleep (it uses
> mutexes) so it cannot be called in preemption-disabled context, and when
> you try to do this, this is the cras
On Sun, Nov 27, 2016 at 9:07 PM, Rick Payne wrote:
>
> I’ve got an application that I’m trying to run under OSv and I was hitting
> an issue where it calls shmget() for an object which doesn’t exist (testing
> to see if its there from a previous run, which of course under OSv it won’t
> be). IPC_
> On 28 Nov 2016, at 10:57, Nadav Har'El wrote:
>
> However, our implementation of the symbol resolution can sleep (it uses
> mutexes) so it cannot be called in preemption-disabled context, and when you
> try to do this, this is the crash you see.
Thanks for the explanation - makes more sense
On Sun, Nov 27, 2016 at 11:23 PM, Rick Payne wrote:
>
> Continuing on with the erlang hacking, I’m running into an assert here:
>
> Assertion failed: sched::preemptable() (core/elf.cc: resolve_pltgot: 664)
>
> [backtrace]
> 0x0022d39e
> 0x0022d3ff <__assert_fail+58>
> 0x0
On Sun, Nov 27, 2016 at 8:16 PM, Waldemar Kozaczuk
wrote:
> Modified openjdk7, openjdk8-fedora, openjdk9-ea-java-base to rely on
> 'java' capability which can be
> provided during build process by adding either java-isolated or
> java-non-isolated to the list of modules (see examples below).
> Th
On 11/27/2016 11:23 PM, Rick Payne wrote:
Continuing on with the erlang hacking, I’m running into an assert here:
Assertion failed: sched::preemptable() (core/elf.cc: resolve_pltgot: 664)
[backtrace]
0x0022d39e
0x0022d3ff <__assert_fail+58>
0x004003c0
0x00404
Hi,
I just submitted 2 patches that are addressing moving java stuff from java/
to modules/java-*. There is still some work to do as far as moving
jolokia-agent and cloudius modules from java-base and however I am not sure
what is the role of each module listed below (apart from jolokia) to know
i
24 matches
Mail list logo