em(4): Adjust RDTR/RADV comment that they should not be used

em(4): Don't configure RADV (default is 0), which is deprecated.

- In 82573 workaround, we still need to configure RADV, since RDTR is used.
- Revoke tunable/sysctl node for RADV.

em(4): Revoke tunable/sysctl node for RDTR, which is dangerous and deprecated.

em(4): Reduce TX interrupt rate and device's TX desc write requests

The RS (report status) bit in the TX desc controls whether DD bit
should be set by device (via write request) and whether TX inter

em(4): Reduce TX interrupt rate and device's TX desc write requests

The RS (report status) bit in the TX desc controls whether DD bit
should be set by device (via write request) and whether TX interrupt
should be generated. By setting RS bit in the last TX desc of
int_tx_nsegs TX descs, we greatly reduce the TX interrupt rate
(from 20000/s to 1200/s for full speed 1472bytes UDP datagrams) and
the number of device's TX desc write requests. This also gives me
additional +10Kpps on 82573E_IAMT. Add sysctl node for int_tx_nsegs,
its default value is 1/16 number of TX descs. The implementation
details are commented near struct adapter's related fields.

show more ...

em(4): Avoid allocating a csum offloading TX desc whenever possible.

According to Intel's PCIe GbE Controllers Open Source Software
Developer's Manual Revision 1.8: a csum offloading TX desc will
pr

em(4): Avoid allocating a csum offloading TX desc whenever possible.

According to Intel's PCIe GbE Controllers Open Source Software
Developer's Manual Revision 1.8: a csum offloading TX desc will
prevent TX data read requests from being pipelined, thus reduce TX
performance. The pipelining effect is not obvious when transmitting
bulk data (e.g. 1472 bytes UDP datagram), but it could be dominant
when transmitting tiny packets. So we should avoid allocating a
csum offloading TX desc whenever possible to take advantage of the
pipelining effect.

On 82573E_IAMT,
Before this commit: ~700Kpps
After this commit: ~990Kpps

The funny thing about this commit is:
Old driver code from Intel's FreeBSD driver 6.2.9 roughly did what
we are doing in this commit, while Intel's FreeBSD driver 6.9.6
simply follows Linux's way to flush the performance to the toilet ...

show more ...

em(4): On TX path, make sure that ether header and ip.ip_hl are contiguous.

This must be done mainly because if hardware TX csum offloading is enabled,
setting up context TX descriptor will need to

em(4): On TX path, make sure that ether header and ip.ip_hl are contiguous.

This must be done mainly because if hardware TX csum offloading is enabled,
setting up context TX descriptor will need to access ip.ip_hl.

show more ...

em(4): Sync with Intel's em-6.9.6(FreeBSD)

In addition to adding support for some chips (e.g. 82574L), this also gives
me the chance to rearrange and clean up if_em.[ch] :)

Noticable changes to the

em(4): Sync with Intel's em-6.9.6(FreeBSD)

In addition to adding support for some chips (e.g. 82574L), this also gives
me the chance to rearrange and clean up if_em.[ch] :)

Noticable changes to the FreeBSD driver:
- The hardware abstraction layer is put into a seperate module (ig_hal)
- IP csum offloading is supported when hardware TX csum offloading is enabled
- mbufs on RX/TX ring are freed in em_stop(), i.e. during "ifconfig emX down"
- TX pattern is adjusted that we test number of avaiable TX descs before
dequeuing mbuf from ifq. We also reserve double spare TX descs for 82544
cards on PCI-X bus, so we will not need to unload the loaded mbuf in the
mid way due to short of TX descs; this at least makes the logic a little
bit simpler.

show more ...

Don't let PROMISC interfere hardware vlan tagging operation. Enable hardware
vlan tagging after this change.

Sync with Intel's em-6.2.9
- Support more chips.
- Implement suspend/resume device methods.
- Correct promisc mode support.
- Code and comment clean up.
- Avoid reinitializing the hardware if only IP

Sync with Intel's em-6.2.9
- Support more chips.
- Implement suspend/resume device methods.
- Correct promisc mode support.
- Code and comment clean up.
- Avoid reinitializing the hardware if only IP address is changed. This is
implemented in a different way from Intel's:
If IFF_RUNNING is set, em_init() will return immediately. So for some
situation in which em_init() must run, IFF_RUNNING is cleared before calling
em_init().
- Rework TX processing:
In em_encap(), save the index of the packet's last TX descriptor (EOP) in
the packet's first descriptor. In em_txeof(), test E1000_TXD_STAT_DD bit
only for last TX descriptor (EOP) of certain packet. If it is set, then
reap the TX descriptors of the packet in an inner loop.
"This change is important for future chips, where DD bit is going
to be set only on the EOP descriptors." -- Jack Vogel at Intel

show more ...

Rename printf -> kprintf in sys/ and add some defines where necessary
(files which are used in userland, too).

Do a major clean-up of the BUSDMA architecture. A large number of
essentially machine-independant drivers use the structures and definitions
in machine-dependant directories that are really machine-

Do a major clean-up of the BUSDMA architecture. A large number of
essentially machine-independant drivers use the structures and definitions
in machine-dependant directories that are really machine-independant in
nature.

Split <machine/bus_dma.h> into machine-depdendant and machine-independant parts
and make the primary access run through <sys/bus_dma.h>.

Remove <machine/bus.h>, <machine/bus_memio.h> and <machine/bus_pio.h>. The
optimizations related to bus_memio.h and bus_pio.h made a huge mess,
introduced machine-specific knowledge into essentially machine-independant
drivers, and required specific #include file orderings to do their job.
They may be reintroduced in some other form later on.

Move <machine/resource.h> to <sys/bus_resource.h>. The contents of the file
is machine-independant or can be made a superset across many platforms.

Make <sys/bus.h> include <sys/bus_dma.h> and <sys/bus_resource.h> and
include <sys/bus.h> where necessary. Remove all #include's of
<machine/resource.h> and <machine/bus.h>. That is, make the BUSDMA
infrastructure integral to I/O-mapped and memory-mapped accesses to devices
and remove a large chunk of machine-specific dependancies from drivers.
bus_if.h and device_if.h are now required to be present when using <sys/bus.h>.

show more ...

Sync with Intel's em(4) driver version 6.1.4
- Add support for new chipset revisions embedded in the ESB2 and ICH8 core
logic chipsets.
- Add PCI id for the PCIe quad port copper adapter. [*]
- Imp

Sync with Intel's em(4) driver version 6.1.4
- Add support for new chipset revisions embedded in the ESB2 and ICH8 core
logic chipsets.
- Add PCI id for the PCIe quad port copper adapter. [*]
- Improve media support for fiber adapters. [*]
- Instead of creating TX DMA map in em_encap(), create all of the during
em_init().
Fix a bug in Intel's code:
The TX desc used to hold mbuf, does not necessarily have the DMA map
that is used to load the mbuf, so in em_clean_transmit_interrupts() or
during em_stop(), wrong DMA map may be used.
- Update README

Bug fixes:
- Properly serialize em_shutdown()
- (BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD) -> BUS_DMASYNC_PREWRITE
- Correct polling(4) support: disable interrupt if polling(4) is enabled.

Tested with an 82540EM.

# I have committed a slightly different version(6.1.4) than the version(6.0.5)
# posted to submit@, since 6.1.4 came immediately after my submission for
# testing/reviewing, and the difference between it and the submitted version
# is really very small (only added [*] part).

show more ...

Remove OLDBRIDGE

Add KTR logging for IF_EM to measure interrupt overhead and packet
receive / transmit / transmit-cleanup operations.

Make all network interrupt service routines MPSAFE part 1/3.

Replace the critical section that was previously used to serialize access
with the LWKT serializer. Integrate the serializer into the IF

Make all network interrupt service routines MPSAFE part 1/3.

Replace the critical section that was previously used to serialize access
with the LWKT serializer. Integrate the serializer into the IFNET structure.
Note that kern.intr_mpsafe must be set to 1 for network interrupts to actually
run MPSAFE. Also note that any interrupts shared with othre non-MP drivers
will cause all drivers on that interrupt to run with the Big Giant Lock.

Network interrupt - Each network driver then simply passes that serializer
to bus_setup_intr() so only a single serializer is required to process the
entire interrupt path. LWKT serialization support is already 100% integrated
into the interrupt subsystem so it will already be held as of when the
registered interrupt procedure is called.

Ioctl and if_* functions - All callers of if_* functions (such as if_start,
if_ioctl, etc) now obtain the IFNET serializer before making the call. Thus
all of these entry points into the driver will now be serialized.

if_input - All code that calls if_input now ensures that the serializer
is held. It will either already be held (when called from a driver), or
the serializer will be wrapped around the call. When packets are forwarded
or bridged between interfaces, the target interface serializer will be
dropped temporarily to avoid a deadlock.

Device Driver access - dev_* entry points into certain pseudo-network
devices now obtain and release the serializer. This had to be done on
a device-by-device basis (but there are only a few such devices).

Thanks to several people for helping test the patch, in particular
Sepherosa Ziehau.

show more ...

Sync with Intel's em(4) driver version 3.2.15
- Add support for 82571EB, 82572EI and 82573E adapters
- Nuke DBG_STATS related code and struct adapter's fields
- Flow control water marks are now calcu

Sync with Intel's em(4) driver version 3.2.15
- Add support for 82571EB, 82572EI and 82573E adapters
- Nuke DBG_STATS related code and struct adapter's fields
- Flow control water marks are now calculated from hardware's register value
instead of using hardwired values(FC_DEFAULT_{LOW, HIGH}_THRESH)
- Flow control pause time is bumped to 512us (old value is 131us) (*)
- In em_detach(), call bus_teardown_intr() if adapter.int_handler_tag != NULL
- Update Intel's LICENSE and README
- Update README to reflect recent changes to device polling
- Minor style changes

(*) Original code misuses 0x1000 (should be 1000) as em_hw.fc_pause_time

show more ...

Sync with FreeBSD dev/em

if_em.c rev 1.2.2.{20,21,22} 1.47 1.63
if_em.h rev 1.1.2.17
if_em_hw.c rev 1.1.2.12
if_em_hw.h rev 1.1.2.12

- Add support for Intel PRO/1000 GT Desktop Adapter
- Add supp

Sync with FreeBSD dev/em

if_em.c rev 1.2.2.{20,21,22} 1.47 1.63
if_em.h rev 1.1.2.17
if_em_hw.c rev 1.1.2.12
if_em_hw.h rev 1.1.2.12

- Add support for Intel PRO/1000 GT Desktop Adapter
- Add support for Intel 82546GB dual port PCIE adapter
- Add fix for 82547 which corrects an issue with Jumbo frames larger than 10k
- Add fix for vlan tagged frames not being properly bridged
- Correct TBI workaround
- Correct LED operation issues
- Do not count RNBC (internal buffer full) towards the RX error count,
since it's not really an error
- Correct a workaround that should only be applied to one adapter. Workaround
was causing device hangs when incorrectly applied to other adapters (*)
- Run em_local_timer() once per second instead of running it once per 2 seconds.
This makes gathering of stats more precise

(*) Detailed problem description is in FreeBSD's PR kern/66634

show more ...

Rewrite the polling code. Instead of trying to do fancy polling enablement
from inside the IF interrupt itself, which creates a headache in the code,
simply allow IFF_POLLING to be set and cleared v

Rewrite the polling code. Instead of trying to do fancy polling enablement
from inside the IF interrupt itself, which creates a headache in the code,
simply allow IFF_POLLING to be set and cleared via ifconfig. This greatly
simplifies both the networking code and the polling code and allows polling
to be enabled and disabled at will on a per-network-interface basis.

* Drivers no longer have to have polling checks in the interrupt path.
* An if_poll function vector has been added. Polling is supported if the
driver initializes the vector.
* Registration command added to the poll function command list.
* Driver code for registration and deregistration is now greatly simplified.

The kernel polling code no longer randomly turns off the polling bit if an
interface goes down or is reset.

Remove IFCAP_POLLING, it serves no purpose.

Fix a couple of bugs in the serializer code. Add a warning in
nexus_setup_intr if a driver tries to specify a serializer and an SPL.
A driver can specify one or the other, not both.

Convert the EM driver to use the new serializer API instead of SPLs.

Add ifconfig poll and ifconfig -poll support to ifconfig, and fix bugs
in the rtsock code that only returned the low 16 bits of the interface
flags so ifconfig properly reports when polling mode is turned on for an
interface.

NOTE to people using polling. You must first enable polling via
kern.polling.enable, and then may specify the 'poll' directive in ifconfig
to enable it on a per interface basis. If IFF_POLLING refuses to be set,
the device does not support polling.

show more ...

Get rid of bus_{disable,enable}_intr(), it wasn't generic enough for
our needs.

Implement some generic atomic.h functions to aid in the implementation of
a low level mutex.

Implement a generic low

Get rid of bus_{disable,enable}_intr(), it wasn't generic enough for
our needs.

Implement some generic atomic.h functions to aid in the implementation of
a low level mutex.

Implement a generic low level sleep-mutex serializer, kern/lwkt_serialize.c.
The serializer is designed to be a replacement for SPL calls but may also
be used for other very low level work (e.g. lockmgr interlocks).

Add a serializer argument to BUS_SETUP_INTR(). When non-NULL, the interrupt
handler will no longer be protected by an SPL so e.g. spl*() will no
longer protect against that device's interrupts.

The IF queueing and dequeueing mechanisms may no longer depend on outside
SPL state because network driver interrupt handlers are no longer required to
enter splnet(). Use critical sections for the moment. The IFQ and
IFF_OACTIVE interactions are not yet MP safe.

show more ...

Add support for adjusting the interrupt throttling rate via
hw.emX.int_throttle_ceil, and change the default from 8000hz
to 10000hz to improve throughput with default TCP window sizes
through a GiGE

Add support for adjusting the interrupt throttling rate via
hw.emX.int_throttle_ceil, and change the default from 8000hz
to 10000hz to improve throughput with default TCP window sizes
through a GiGE switch in a LAN environment.

Note the following test results using iperf between two machines:

* Shuttle Athlon 64 3200+ box, EM card in 32 bit PCI slot
* 2 machines connected through a GiGE switch
* All other hw.em0 delays set to 0 on both sides
* throttle settings set on both sides
* -w option set on iperf client AND server for 63.5KB window
* software interrupt throttling has been turned off for these tests


throttle result result
freq (32.5KB win) (63.5KB win)
(default)
-------- -------------- -----------

maxrate 481 MBit/s 533 MBit/s (not sure what's going on here)
120000 518 MBit/s 558 MBit/s (not sure what's going on here)
100000 613 MBit/s 667 MBit/s (not sure what's going on here)
70000 679 MBit/s 691 MBit/s
60000 668 MBit/s 694 MBit/s
50000 678 MBit/s 684 MBit/s
40000 694 MBit/s 696 MBit/s
30000 694 MBit/s 696 MBit/s
20000 698 MBit/s 703 MBit/s
10000 707 MBit/s 716 MBit/s
9000 708 MBit/s 716 MBit/s
8000 710 MBit/s 717 MBit/s <--- drop off pt 32.5KB win
7000 683 MBit/s 716 MBit/s
6000 680 MBit/s 720 MBit/s
5000 652 MBit/s 718 MBit/s <--- drop off pt 63.5KB win
4000 555 Mbit/s 695 MBit/s
3000 522 MBit/s 533 MBit/s <--- GiGE throttling likely
2000 449 MBit/s 384 MBit/s (256 ring descriptors =
1000 260 MBit/s 193 MBit/s 2500 hz minimum)

show more ...

Continue cleaning em(4).

Merge PCI resource allocation / deallocation into em_attach and em_detach.
Change the order of allocations in em_attach to work without splimp()
protection.
Remove adapter->

Continue cleaning em(4).

Merge PCI resource allocation / deallocation into em_attach and em_detach.
Change the order of allocations in em_attach to work without splimp()
protection.
Remove adapter->unit, adapter->next and adapter->prev.
Use em_detach for the error handling in em_attach as well.
Move the reset of if_timer into em_stop.
Use adapter->hw.mac_addr instead of arpcom's entry.

show more ...

Sync em(4) with FreeBSD current. Most important is the initial bus DMA support.
Also use callout directly and not timeout.

Change WAIT_FOR_AUTO_NEG_DEFAULT to 0. Do not wait for auto-negotiation to
complete before continuing. if_em can take over two seconds to autoneg, and it
renegotiates virtually every time the inter

Change WAIT_FOR_AUTO_NEG_DEFAULT to 0. Do not wait for auto-negotiation to
complete before continuing. if_em can take over two seconds to autoneg, and it
renegotiates virtually every time the interface is manipulated. For example,
running dhclient will cause if_em to renegotiate 3 times in a row.

Note that even when we were waiting for autoneg, the interface would not
always be completely up after the autoneg completed. This is possibly
due to the watchdog timer (which is on a 2 second interval). In fact, on
my test box this is what was tickling the recently fixed DNS bug (timing
out on part of the dns search path) and subsequent NFS mount failures.

show more ...

Update the 825xx GigE support. Add a large number of new device id's and
support for the 82541 (e.g. Intel PRO/1000 MT DESKTOP ADAPTER) and 82544.
Also add sysctl support for various interrupt relat

Update the 825xx GigE support. Add a large number of new device id's and
support for the 82541 (e.g. Intel PRO/1000 MT DESKTOP ADAPTER) and 82544.
Also add sysctl support for various interrupt related delay parameters,
and add workarounds for chip issues.

In-Discussion-With: "Peter Avalos" <pavalos@theshell.com>
Submitted-by: Hiten Pandya <hmp@backplane.com>
Adapted-from: FreeBSD-5

show more ...

kernel tree reorganization stage 1: Major cvs repository work (not logged as
commits) plus a major reworking of the #include's to accomodate the
relocations.

* CVS repository files manually move

kernel tree reorganization stage 1: Major cvs repository work (not logged as
commits) plus a major reworking of the #include's to accomodate the
relocations.

* CVS repository files manually moved. Old directories left intact
and empty (temporary).

* Reorganize all filesystems into vfs/, most devices into dev/,
sub-divide devices by function.

* Begin to move device-specific architecture files to the device
subdirs rather then throwing them all into, e.g. i386/include

* Reorganize files related to system busses, placing the related code
in a new bus/ directory. Also move cam to bus/cam though this may
not have been the best idea in retrospect.

* Reorganize emulation code and place it in a new emulation/ directory.

* Remove the -I- compiler option in order to allow #include file
localization, rename all config generated X.h files to use_X.h to
clean up the conflicts.

* Remove /usr/src/include (or /usr/include) dependancies during the
kernel build, beyond what is normally needed to compile helper
programs.

* Make config create 'machine' softlinks for architecture specific
directories outside of the standard <arch>/include.

* Bump the config rev.

WARNING! after this commit /usr/include and /usr/src/sys/compile/*
should be regenerated from scratch.

show more ...