- Jan 14, 2013
- Jan 13, 2013
-
-
Avi Kivity authored
-
Avi Kivity authored
Required by pthread_getattr_np(), in turn required by the jvm.
-
Avi Kivity authored
-
Dor Laor authored
* 'master' of github.com:/avikivity/osv: pthread: implement pthread_create()
-
Dor Laor authored
queue size correctly.
-
Dor Laor authored
-
Avi Kivity authored
-
- Jan 12, 2013
-
-
Guy Zana authored
-
- Jan 11, 2013
-
-
Avi Kivity authored
Apparently glibc defines it as a macro, so it doesn't compile. #undef it back.
-
Avi Kivity authored
Caught by make mode=release.
-
Avi Kivity authored
-
Avi Kivity authored
-
Avi Kivity authored
Otherwise 'make' deletes the intermediate .o file and causes spurious rebuilds.
-
Avi Kivity authored
-
Avi Kivity authored
Break long lines
-
Avi Kivity authored
Cuts boot time dramatically
-
Avi Kivity authored
-
Christoph Hellwig authored
-
Christoph Hellwig authored
-
- Jan 10, 2013
-
-
Avi Kivity authored
-
Avi Kivity authored
Instead of using the variable's value, we used its name. There isn't even a name for this kind of bug.
-
Avi Kivity authored
-
Avi Kivity authored
-
Avi Kivity authored
Since file::file() did not open the file, file::~file() must not free it. Use ::dup() so we operate on a clone.
-
Avi Kivity authored
-
Avi Kivity authored
-
Avi Kivity authored
-
Avi Kivity authored
-
Avi Kivity authored
Do nothing yet (used for CLOEXEC, meaningless for us).
-
Avi Kivity authored
-
Avi Kivity authored
Conflicts: libc/fd.cc __xstat change: obsoleted by upstream open64(): moved to fs/vfs/main.c
-
Avi Kivity authored
-
Avi Kivity authored
When we aligned the sizes for mmap(), we neglected to clear memory between the end of the read data and the beginning of anonymous memory. If a variable sat there, it would not be initialized correctly.
-
Avi Kivity authored
Could be better, but waiting for vfs revamp anyway.
-
Christoph Hellwig authored
-
Christoph Hellwig authored
This seems to break the actual jvm execution for some reason.
-
Christoph Hellwig authored
Also based on prex code.
-