This patch fixes part of TCP instability/low throughput problems on Xen.

On transmit networking stack passes to Xen netfront driver a list of mbufs
of up to PAGE_SIZE (4096) bytes each.
In case the list consists of more than MAX_TX_REQ_FRAGS (18) fragments
driver tries to defragment it with m_defrag().

m_defrag() in turn tries to build list of buffers of up to MCLBYTES
(2048) bytes each. This leads to even longer chain and packet being dropped.
As a result around 1% of TCP segments are lost and intensive TCP
retransmissions and slowdowns occur.

This patch makes m_defrag() use Jumbo packets zone for allocation, i.e.
PAGE_SIZE per mbuf.

This patch is pretty similar to what Glauber done on RX path before.

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>

Add a 'make check' target that runs whitelisted tests under OSv.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Like commit b17819df ("mmu: optimize searching for vmas") but for rest of
core/mmu.cc.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Make the adding file message less intimidating.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>

We should pass the image path to run.py

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>

This will be needed my mkzfs.py

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>

Start listening as soon as the guest is running. This makes host
mkzfs.py conneciton less racy and fixes OSv image creation problems I'm
seeing.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

strlcpy() expects buffer length not string length.

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>

When the build/ directory is on tmpfs then 'cache=none' option makes
qemu fail with error:

    could not open disk image build/release/usr.img : Invalid argument

Using tmpfs for build/ makes a _huge_ difference in build time on some
setups.

This patch adds fallback to 'unsafe' mode when direct IO is not
supported.

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>

Not supported by older qemus.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Not supported on osv, and reported to break the build on some systems.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Switch osv to built-in mkfs rather than zfs-fuse.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Drop sudo and global /zfs temporary directory.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

This drops the superuser requirement from the build process, and reduces
the build prerequisite list.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Implement a parser for the cpio "newc" format.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Useful when a filesystem is not yet available.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

This allows using the boost libraries in shared objects.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Dump registers on general protection fault for debugging purposes.  Even
if you have gdb available, getting to the exception frame is not always
possible after OSv has crashed.

Example output looks as follows:

registers:
RIP: 0x0000100000b7e913  RFL: 0x0000000000010202  CS:  0x0000000000000008  SS:  0x0000000000000010
RAX: 0xffffc000418ed278  RBX: 0xffffc00041b2c050  RCX: 0x0000000000000004  RDX: 0x0000000000000000
RSI: 0x0000000000000001  RDI: 0x43e0000000000000  RBP: 0x0000200008548d10  R8:  0xffffc000426e3010
R9:  0x0000000000000004  R10: 0x43e0000000000000  R11: 0xffffc00041b2c050  R12: 0xffffc000418ed1e8
R13: 0x0000000000000004  R14: 0x43e0000000000000  R15: 0xffffc00041b2c050  RSP: 0x0000200008548aa0
general protection fault

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Make the two definitions exactly equal so the compiler does not complain.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Create a file system on top of the pool.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

zfs utilities want /etc/mnttab, let them have it.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Ignoring EPOLLET will generate more wakeups, but should work with most
applications.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

For the moment, it only creates a zpool, not a filesystem.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Large partitions set head=254, not 255.  Also be more accurate calculating
the exact point of overflow.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

More realistic.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

It's a superset of epoll_create(), so implement the latter in terms of the
former.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Needed by boost.asio.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

std::streambuf<char> is defined as an "extern template", which means it is
instantiated in the library, not at the point-of-use, in order to reduce
object size and compilation time.

One of its member functions, std::basic_streambuf::seekpos(...) has a
parameter whose type depends on mbstate_t; this embeds the type name into
the mangled function name.

On Linux, mbstate_t is a typedef for __mbstate_t, so the latter type name
is embedded into the mangled function name.  On osv, there is no such typedef,
so mbstate_t is embedded.  This causes lookup to fail when loading a shared
object that uses streambufs.

Fix by defining mbstate_t in the same way as glibc.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>