[PATCH] S2io: Large Receive Offload (LRO) feature(v2) for Neterion (s2io) 10GbE Xframe PCI-X and PCI-E NICs

Hi,
Below is a patch for the Large Receive Offload feature.
Please review and let us know your comments.

LRO algorithm was described in an OLS 2005 presentation, located at
ftp.s2io.com
user: linuxdocs
password: HALdocs

The same ftp site has Programming Manual for Xframe-I ASIC.
LRO feature is supported on Neterion Xframe-I, Xframe-II and
Xframe-Express 10GbE NICs.

Brief description:
The Large Receive Offload(LRO) feature is a stateless offload
that is complementary to TSO feature but on the receive path.
The idea is to combine and collapse(upto 64K maximum) in the
driver, in-sequence TCP packets belonging to the same session.
It is mainly designed to improve 1500 mtu receive performance,
since Jumbo frame performance is already close to 10GbE line
rate. Some performance numbers are attached below.

Implementation details:
1. Handle packet chains from multiple sessions(current default
MAX_LRO_SESSSIONS=32).
2. Examine each packet for eligiblity to aggregate. A packet is
considered eligible if it meets all the below criteria.
  a. It is a TCP/IP packet and L2 type is not LLC or SNAP.
  b. The packet has no checksum errors(L3 and L4).
  c. There are no IP options. The only TCP option supported is timestamps.
  d. Search and locate the LRO object corresponding to this
     socket and ensure packet is in TCP sequence.
  e. It's not a special packet(SYN, FIN, RST, URG, PSH etc. flags are not set).
  f. TCP payload is non-zero(It's not a pure ACK).
  g. It's not an IP-fragmented packet.
3. If a packet is found eligible, the LRO object is updated with
   information such as next sequence number expected, current length
   of aggregated packet and so on. If not eligible or max packets
   reached, update IP and TCP headers of first packet in the chain
   and pass it up to stack.
4. The frag_list in skb structure is used to chain packets into one
   large packet.

Kernel changes required: None

Performance results:
Main focus of the initial testing was on 1500 mtu receiver, since this
is a bottleneck not covered by the existing stateless offloads.

There are couple disclaimers about the performance results below:
1. Your mileage will vary!!!! We initially concentrated on couple pci-x
2.0 platforms that are powerful enough to push 10 GbE NIC and do not
have bottlenecks other than cpu%;  testing on other platforms is still
in progress. On some lower end systems we are seeing lower gains.

2. Current LRO implementation is still (for the most part) software based,
and therefore performance potential of the feature is far from being realized.
Full hw implementation of LRO is expected in the next version of Xframe ASIC.

Performance delta(with MTU=1500) going from LRO disabled to enabled:
IBM 2-way Xeon (x366) : 3.5 to 7.1 Gbps
2-way Opteron : 4.5 to 6.1 Gbps
Signed-off-by: NRavinandan Arakali <ravinandan.arakali@neterion.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

	There is a problem with fragmented skb in s2io driver version 2.0.9.4
	available in 2.6.16-rc1 kernel. The adapter will fail to transmit if
	any scatter-gather skb arrives.  This patch provides fix for the above
	described problem.
Signed-off-by: NAnanda Raju <ananda.raju@neterion.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

On my laptop, the b44 device is created and the carrier state defaults
to ON when created by alloc_etherdev. This means tools like NetworkManager
see the carrier as On and try and bring the device up.  The correct thing
to do is mark the carrier as Off when device is created.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

tmp in ace_init is u32 thus rendering read_eeprom_byte() return values
checks useless.
Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Since get_settings() returns a signed int and it gets checked
for < 0 to catch an error, res should be a signed int too.
Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The compat layer timeout handling changes in:

9f72949f

are busted.  This is most easily seen with an X application
that uses sub-second select/poll timeout such as emacs.  You
hit a key and it takes a second or so before the app responds.

The two ROUND_UP() calls upon entry are using {tv,ts}_sec where it
should instead be using {tv_usec,ts_nsec}, which perfectly explains
the observed incorrect behavior.

Another bug shot down with git bisect.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Or else we break on ppc32 and other 32-bit platforms.

Based upon a patch from Harald Welte.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix an unnecessary softlockup watchdog warning in the ia64
uncached_build_memmap() that occurs occasionally at 256p and always at
512p.  The problem occurs at boot time.
Signed-off-by: NJohn Hawkes <hawkes@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NTony Luck <tony.luck@intel.com>

Migrate perfmon from using an old semaphore to a completion handler.
Signed-off-by: NJes Sorensen <jes@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NTony Luck <tony.luck@intel.com>

Migrate arch/ia64/ia32/sys_ia32 to using a mutex for mmap protection.
Signed-off-by: NJes Sorensen <jes@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NTony Luck <tony.luck@intel.com>

This also includes by necessity _TIF_RESTORE_SIGMASK support,
which actually resulted in a lot of cleanups.

The sparc signal handling code is quite a mess and I should
clean it up some day.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

some driver clean ups, and a re-posting of changes that are needed

to match the updated TPS.
Signed-off-by: NMark Gross <mark.gross@intel.com>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This is a subset of the bluesmoke project core code, stripped of the NMI work
which isn't ready to merge and some of the "interesting" proc functionality
that needs reworking or just has no place in kernel.  It requires no core
kernel changes except the added scrub functions already posted.

The goal is to merge further functionality only after the core code is
accepted and proven in the base kernel, and only at the point the upstream
extras are really ready to merge.

From: doug thompson <norsk5@xmission.com>

  This converts EDAC to sysfs and is the final chunk neccessary before EDAC
  has a stable user space API and can be considered for submission into the
  base kernel.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Signed-off-by: NJesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Ndoug thompson <norsk5@xmission.com>
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

EDAC requires a way to scrub memory if an ECC error is found and the chipset
does not do the work automatically.  That means rewriting memory locations
atomically with respect to all CPUs _and_ bus masters.  That means we can't
use atomic_add(foo, 0) as it gets optimised for non-SMP

This adds a function to include/asm-foo/atomic.h for the platforms currently
supported which implements a scrub of a mapped block.

It also adjusts a few other files include order where atomic.h is included
before types.h as this now causes an error as atomic_scrub uses u32.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add the sys_pselect6() and sys_poll() calls to the i386 syscall table.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The following implementation of ppoll() and pselect() system calls
depends on the architecture providing a TIF_RESTORE_SIGMASK flag in the
thread_info.

These system calls have to change the signal mask during their
operation, and signal handlers must be invoked using the new, temporary
signal mask. The old signal mask must be restored either upon successful
exit from the system call, or upon returning from the invoked signal
handler if the system call is interrupted. We can't simply restore the
original signal mask and return to userspace, since the restored signal
mask may actually block the signal which interrupted the system call.

The TIF_RESTORE_SIGMASK flag deals with this by causing the syscall exit
path to trap into do_signal() just as TIF_SIGPENDING does, and by
causing do_signal() to use the saved signal mask instead of the current
signal mask when setting up the stack frame for the signal handler -- or
by causing do_signal() to simply restore the saved signal mask in the
case where there is no handler to be invoked.

The first patch implements the sys_pselect() and sys_ppoll() system
calls, which are present only if TIF_RESTORE_SIGMASK is defined. That
#ifdef should go away in time when all architectures have implemented
it. The second patch implements TIF_RESTORE_SIGMASK for the PowerPC
kernel (in the -mm tree), and the third patch then removes the
arch-specific implementations of sys_rt_sigsuspend() and replaces them
with generic versions using the same trick.

The fourth and fifth patches, provided by David Howells, implement
TIF_RESTORE_SIGMASK for FR-V and i386 respectively, and the sixth patch
adds the syscalls to the i386 syscall table.

This patch:

Add the pselect() and ppoll() system calls, providing core routines usable by
the original select() and poll() system calls and also the new calls (with
their semantics w.r.t timeouts).
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Use the generic sys_rt_sigsuspend.
Signed-off-by: NJeff Dike <jdike@addtoit.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add support for TIF_RESTORE_SIGMASK.  I copy the i386 handling of the flag.
sys_sigsuspend is also changed to follow i386.
Also a bit of cleanup -
   turn an if into a switch
   get rid of a couple more emacs formatting comments
Signed-off-by: NJeff Dike <jdike@addtoit.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Implement the TIF_RESTORE_SIGMASK flag in the new arch/powerpc kernel, for
both 32-bit and 64-bit system call paths.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Handle TIF_RESTORE_SIGMASK as added by David Woodhouse's patch entitled:

        [PATCH] 2/3 Add TIF_RESTORE_SIGMASK support for arch/powerpc
        [PATCH] 3/3 Generic sys_rt_sigsuspend

It does the following:

 (1) Declares TIF_RESTORE_SIGMASK for i386.

 (2) Invokes it over to do_signal() when TIF_RESTORE_SIGMASK is set.

 (3) Makes do_signal() support TIF_RESTORE_SIGMASK, using the signal mask saved
     in current->saved_sigmask.

 (4) Discards sys_rt_sigsuspend() from the arch, using the generic one instead.

 (5) Makes sys_sigsuspend() save the signal mask and set TIF_RESTORE_SIGMASK
     rather than attempting to fudge the return registers.

 (6) Makes sys_sigsuspend() return -ERESTARTNOHAND rather than looping
     intrinsically.

 (7) Makes setup_frame(), setup_rt_frame() and handle_signal() return 0 or
     -EFAULT rather than true/false to be consistent with the rest of the
     kernel.

Due to the fact do_signal() is then only called from one place:

 (8) Makes do_signal() no longer have a return value is it was just being
     ignored; force_sig() takes care of this.

 (9) Discards the old sigmask argument to do_signal() as it's no longer
     necessary.

(10) Makes do_signal() static.

(11) Marks the second argument to do_notify_resume() as unused. The unused
     argument should remain in the middle as the arguments are passed in as
     registers, and the ordering is specific in entry.S

Given the way do_signal() is now no longer called from sys_{,rt_}sigsuspend(),
they no longer need access to the exception frame, and so can just take
arguments normally.

This patch depends on sys_rt_sigsuspend patch.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Handle TIF_RESTORE_SIGMASK as added by David Woodhouse's patch entitled:

        [PATCH] 2/3 Add TIF_RESTORE_SIGMASK support for arch/powerpc
        [PATCH] 3/3 Generic sys_rt_sigsuspend

It does the following:

 (1) Declares TIF_RESTORE_SIGMASK for FRV.

 (2) Invokes it over to do_signal() when TIF_RESTORE_SIGMASK is set.

 (3) Makes do_signal() support TIF_RESTORE_SIGMASK, using the signal mask saved
     in current->saved_sigmask.

 (4) Discards sys_rt_sigsuspend() from the arch, using the generic one instead.

 (5) Makes sys_sigsuspend() save the signal mask and set TIF_RESTORE_SIGMASK
     rather than attempting to fudge the return registers.

 (6) Makes sys_sigsuspend() return -ERESTARTNOHAND rather than looping
     intrinsically.

 (7) Makes setup_frame(), setup_rt_frame() and handle_signal() return 0 or
     -EFAULT rather than true/false to be consistent with the rest of the
      kernel.

Due to the fact do_signal() is then only called from one place:

 (8) Make do_signal() no longer have a return value is it was just being
     ignored; force_sig() takes care of this.

 (9) Discards the old sigmask argument to do_signal() as it's no longer
     necessary.

This patch depends on the FRV signalling patches as well as the
sys_rt_sigsuspend patch.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The TIF_RESTORE_SIGMASK flag allows us to have a generic implementation of
sys_rt_sigsuspend() instead of duplicating it for each architecture.  This
provides such an implementation and makes arch/powerpc use it.

It also tidies up the ppc32 sys_sigsuspend() to use TIF_RESTORE_SIGMASK.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Wire up the x86_64 syscalls.
Signed-off-by: NUlrich Drepper <drepper@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Wire up the x86 syscalls
Signed-off-by: NUlrich Drepper <drepper@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Here is a series of patches which introduce in total 13 new system calls
which take a file descriptor/filename pair instead of a single file
name.  These functions, openat etc, have been discussed on numerous
occasions.  They are needed to implement race-free filesystem traversal,
they are necessary to implement a virtual per-thread current working
directory (think multi-threaded backup software), etc.

We have in glibc today implementations of the interfaces which use the
/proc/self/fd magic.  But this code is rather expensive.  Here are some
results (similar to what Jim Meyering posted before).

The test creates a deep directory hierarchy on a tmpfs filesystem.  Then
rm -fr is used to remove all directories.  Without syscall support I get
this:

real    0m31.921s
user    0m0.688s
sys     0m31.234s

With syscall support the results are much better:

real    0m20.699s
user    0m0.536s
sys     0m20.149s

The interfaces are for obvious reasons currently not much used.  But they'll
be used.  coreutils (and Jeff's posixutils) are already using them.
Furthermore, code like ftw/fts in libc (maybe even glob) will also start using
them.  I expect a patch to make follow soon.  Every program which is walking
the filesystem tree will benefit.
Signed-off-by: NUlrich Drepper <drepper@redhat.com>
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@ftp.linux.org.uk>
Acked-by: NIngo Molnar <mingo@elte.hu>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

find_exported_dentry contains two duplicate loops to find an alias that the
acceptable callback likes.  Split this out to a new helper and switch from
list_for_each to list_for_each_entry to make it more readable.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NNeilBrown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

A recent patch which checked the return status of vfs_getattr in nfsd,
completely missed the nfsproc.c (NFSv2) part.  Here is it.

This patch moved the call to vfs_getattr from the xdr encoding (at which point
it is too late to return an error) to the call handling.  This means several
calls to vfs_getattr are needed in nfsproc.c.  Many are encapsulated in
nfsd_return_attrs and nfsd_return_dirop.
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

nfsd_sync* return an errno, which usually needs to be converted to an errno,
sometimes immediately, sometimes a little later.

Also, nfsd_setattr returns an nfserr which SHOULDN'T be converted from
an errno (because it isn't one).

Also some tidyups of the form:
  err = XX
  err = nfserrno(err)
and
  err = XX
  if (err)
      err = nfserrno(err)
become
  err = nfserrno(XX)
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

missing nfserrno() in default case of a switch by return value of
posix_lock_file(); as the result we send negative host-endian to clients that
expect positive network-endian, preferably mentioned in RFC...  BTW, that case
is not impossible - posix_lock_file() can return -ENOLCK and we do not handle
that one explicitly.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

->rp_status is network-endian and nobody byteswaps it before sending to
client; putting NFSERR_SERVERFAULT instead of nfserr_serverfault in there is
not nice...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>