| Age | Commit message (Collapse) | Author |
|---|
|
- Removed receive buffer replenishment tasklet s2io_tasklet and instead
allocating the receive buffers in either the interrupt handler (no napi)
or the napi handler (napi enabled).
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
- Resubmit #3
- Added s2io_vlan_rx_kill_vid entry point function for unregistering vlan.
- Fix to aggregate vlan packets. IP offset is incremented by
4 bytes if the packet contains vlan header.
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Resubmit #2
- Transmit fifo selection based on TCP/UDP ports.
- Added tx_steering_type loadable parameter for transmit fifo selection.
0x0 NO_STEERING: Default FIFO is selected.
0x1 TX_PRIORITY_STEERING: FIFO is selected based on skb->priority.
0x2 TX_DEFAULT_STEERING: FIFO is selected based on L4 Ports.
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Resubmit #3
Multiqueue netwrok device support implementation.
- Added a loadable parameter "multiq" to enable/disable multiqueue support,
by default it is disabled.
- skb->queue_mapping is not used for queue/fifo selection. FIFO selection is
based on skb->priority.
- Added per FIFO flags FIFO_QUEUE_START and FIFO_QUEUE_STOP. Check this flag
for starting and stopping netif queue and update the flags accordingly.
- In tx_intr_handler added a check to ensure that we have free TXDs before wak-
ing up the queue.
- Added helper functions for queue manipulation(start/stop/wakeup) to invoke
appropriate netif_ functions.
- Calling netif_start/stop for link up/down case respectively.
- As per Andi kleen's review comments, using skb->priority field for FIFO
selection.
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Resubmitting patch from Al Viro <viro@zeniv.linux.org.uk>, with subject -
[PATCH] s2io LRO bugs.
a) initiate_new_session() sets ->tcp_ack to ntohl(...); everything
else stores and expects to find there the net-endian value.
b) check for monotonic timestamps in verify_l3_l4_lro_capable()
compares the value sitting in TCP option (right there in the skb->data,
net-endian 32bit) with the value picked from earlier packet.
Doing that without ntohl() is an interesting idea and it might even
work occasionally; unfortunately, it's quite broken.
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fixes to enable multiple transmit fifos (upto a maximum of eight).
- Moved single tx_lock from struct s2io_nic to struct fifo_info.
- Moved single ufo_in_band_v structure from struct s2io_nic to struct
fifo_info.
- Assign the respective interrupt number for the transmitting fifo in the
transmit descriptor (TXD).
- Added boundary checks for number of FIFOs enabled and FIFO length.
Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
- Support to add/delete/store/restore 64 and 128 Ethernet addresses for Xframe I and Xframe II respectively.
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
s2io, rename BIT macro
BIT macro will be global definiton of (1<<x)
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Ramkrishna Vepa <ram.vepa@neterion.com>
Cc: Rastapur Santosh <santosh.rastapur@neterion.com>
Cc: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Cc: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
- Added set_mac_address driver entry point
- Copying permanent mac address to dev->perm_addr
- Incorporated following review comments from Jeff
- Converted the macro to a function and removed call to memset
- regarding function naming convention, for all callbacks and entry points
will have 's2io_' prefix and helper functions will have 'do_s2io_' prefix.
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Removed bimodal interrupt support - unused feature
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
[also, trim trailing whitespace]
|
|
- Code Optimization of s2io_isr function.
- Isr check using per device napi variable instead of driver global.
- Reduced from 3 to 1 if condition before check for processing packet receive
packets.
- Implemented Jeff's comment to use synchronize_irq. Removed the isr_cnt
variable as it became redundant.
- One time de assert the interrupts by writing all F's to the general_int_mask
register instead of de asserting by clearing the source of interrupts with
multiple writes which causes loss of interrupts (race conditions). It is
entirely possible that before the driver has a chance to mask the asserted
alarm bit, another alarm/traffic interrupt bit gets asserted as well. In
this case Herc will keep the INTA line asserted and the bridge will not
send a new Assert_INTA message upstream.
[ Resolved conflicts due to napi_struct changes... -DaveM ]
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
- Added check to return from the traffic handling function, if the card status
is DOWN.
- Implemented Jeff's comments on incorrect return value in s2io_poll function.
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Removed the unused variable, intr_type, in device private structure.
[ Resolve conflicts with napi_struct changes... -DaveM ]
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
- Added support to poll entire set of device errors and alarams.
- A note on how device errors and alarms are handled:
- The adapter will automatically recover from uncorrectable ECC errors.
Packets containing corrupted data will be dropped (not transmitted) or tagged
as invalid before being passed to the host.
- The adapter cannot recover from any internal state machine errors. A state
machine error requires a device reset.
- Any internal error that could potentially result in .store trampling.
(undesirable PCI behaviour)is tagged as a "serious error". In such cases
the adapter will give up its ability to be a bus master. In this situation
the host will still be able to read internal device registers in order to
generate an error report. A device reset is necessary to return to normal
operation.
- In the event of a pcix data parity error, the adapter will automatically
disable itself. Adapter_En will automatically transition from '1' to '0' and
the adapter will enter its clean-up routine. Once the device has achieved
quiescence, an adapter reset should be performed.
- Replaced alarm_intr_handler() with s2io_handle_errors().
- Added statistic counters to monitor the alarms.
[ Fix warnings wrt. do_s2io_chk_alarm_bit(), Callers pass in an
"unsigned long long *" but the function takes a "u64 *" which is
different on many 64-bit platforms. -DaveM ]
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Added support to unmask entire set of device errors and alarams.
Alarm interrupts are generated for a myriad of purposes, ranging from
illegal operations or requests to internal state machine errors and
uncorrectable data corruption errors. In several cases the adapter can
recover gracefully from unexpected events; however, in some cases, a device
reset may be necessary. This patch handles alarms generated by all the
blocks within the device.
The adapter generates the following types of alarms:
1. Link state transitions (local/remote fault) or other link-related
problems.
2. Problems with any device peripherals, including the EEPROM, FLASH,
etc.
3. Correctable ECC errors (single-bit errors) on internal data
structures or frame data.
4. Uncorrectable ECC errors (multi-bit errors) on internal data
structures or frame data.
5. State machine errors, which indicate that internal control
structures have become corrupted.
6. PCI related errors, including parity errors or illegal transactions.
7. Other unexpected events.
- Implemented Jeff's review comments to use do_s2io_write_bits function to avoid
duplicate codes.
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Calling store_xmsi_data to store the MSI-X datas during initialization
in s2io-init_nic function
- Disabling NAPI when MSI-X is enabled
- Freeing sp->entries and sp->s2io_entries in s2io_rem_isr
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Making MSIX as default intr_type
- Driver will test MSI-X by issuing test MSI-X vector and if fails it will
fallback to INTA
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Several devices have multiple independant RX queues per net
device, and some have a single interrupt doorbell for several
queues.
In either case, it's easier to support layouts like that if the
structure representing the poll is independant from the net
device itself.
The signature of the ->poll() call back goes from:
int foo_poll(struct net_device *dev, int *budget)
to
int foo_poll(struct napi_struct *napi, int budget)
The caller is returned the number of RX packets processed (or
the number of "NAPI credits" consumed if you want to get
abstract). The callee no longer messes around bumping
dev->quota, *budget, etc. because that is all handled in the
caller upon return.
The napi_struct is to be embedded in the device driver private data
structures.
Furthermore, it is the driver's responsibility to disable all NAPI
instances in it's ->stop() device close handler. Since the
napi_struct is privatized into the driver's private data structures,
only the driver knows how to get at all of the napi_struct instances
it may have per-device.
With lots of help and suggestions from Rusty Russell, Roland Dreier,
Michael Chan, Jeff Garzik, and Jamal Hadi Salim.
Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra,
Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan.
[ Ported to current tree and all drivers converted. Integrated
Stephen's follow-on kerneldoc additions, and restored poll_list
handling to the old style to fix mutual exclusion issues. -DaveM ]
Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
- Checking for the return value of pci map function
- Implemented Francois Romieu's comments on eliminating code duplication
using goto
- Implemented Francois Romieu's comments on using a temporary variable for
accessing statistics structure
Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Removed MSI support from driver - unused feature
- Replaced request_mem_region with pci_request_regions
Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Removed 3 buffer mode support from driver - unused feature
- Incorporated Jeff Garzik's comments on elimination of inline typecasting
- Code cleanup : Removed a few extra spaces
Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Instead of all drivers reading pci config space to get the revision
ID, they can now use the pci_device->revision member.
This exposes some issues where drivers where reading a word or a dword
for the revision number, and adding useless error-handling around the
read. Some drivers even just read it for no purpose of all.
In devices where the revision ID is being copied over and used in what
appears to be the equivalent of hotpath, I have left the copy code
and the cached copy as not to influence the driver's performance.
Compile tested with make all{yes,mod}config on x86_64 and i386.
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Acked-by: Dave Jones <davej@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
s2io cleanup suggestions, per discussion on mailing lists.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
This patch adds PCI error recovery support to the
s2io 10-Gigabit ethernet device driver. Third revision,
blocks interrupts and the watchdog.
Tested, seems to work well.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Acked-by: Ramkrishna Vepa <Ramkrishna.Vepa@neterion.com>
Cc: Raghavendra Koushik <raghavendra.koushik@neterion.com>
Cc: Wen Xiong <wenxiong@us.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
1. Added statistics for link up/down, last link up/down.
2. Statistics for memory allocated/freed.
3. Changed level of some DBG_PRINTs.
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
1. Added code to count the number of memory failures.
2. Added watchdog_timer counter.
3. Changed debug level for some DBG_PRINTS.
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Adding get_ringparam ethtool option.
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Remove unused variables from s2io_nic structure
- Changed the memory failure printk messages to print only in debug mode
- Updated the copyright messages
(Resending; due to patch being corrupted)
Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Added code to not to strip vlan tag when driver is in promiscuous mode
- Added module loadable parameter 'vlan_tag_strip" through which user can
enable or disable vlan stripping irrespective of mode
( promiscuous or non-promiscuous ).
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
- Optimized delay to wait for command completion so as to reduce the
initialization wait time.
- Disable differentiated services steering. By default RMAC is configured to
steer traffic with certain DS codes to other queues. Driver must initialize
the DS memory to 0 to make sure that DS steering will not be used by default.
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Removed namespace collisions due to usage of nic_t as per Ralf's patch
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
1. Fix for reset and link handling.
2. Allow for promiscuos mode and multicast state be maintained through
ifconfig up and down.
3. Support to print adapter serial number.
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
This patch adds two load parameters napi and ufo. Previously NAPI was
compilation option with these changes wan enable disable NAPI using load
parameter. Also we are introducing ufo load parameter to enable/disable
ufo feature
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Fix up for make allyesconfig.
Signed-Off-By: David Howells <dhowells@redhat.com>
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
|
|
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
This patch contains some of the bug fixes and enhancements done to
s2io driver. Following are the brief description of changes
1. code cleanup to handle gso modification better
2. Move repeated code in rx path, to a common function
s2io_chk_rx_buffers()
3. Bug fix in MSI interrupt
4. clear statistics when card is down
5. Avoid linked list traversing in lro aggregation.
6. Use pci_dma_sync_single_for_cpu for buffer0 in case of 2/3
buffer mode.
7. ethtool tso get/set functions to set clear NETIF_F_TSO6
8. Stop LRO aggregation when we receive ECN notification
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
This patch contains some of the bug fixes and enhancements done to
s2io driver. Following are the brief description of changes
1. Introduced macro "S2IO_PARM_INT" for declaring integer load parameter
2. UDP_RR test failure, memset txdl after Tx completion
3. PXE boot may leave adapter in unknown state so do reset in probe.
4. Add Tx completion code in netpoll
5. In s2io_vpd_read() move array vpd_data[] to pointer, saves stack memory
6. Fix bug in ethtool online test
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Modification and bug fixes with respect to irq registration.
- Enable interrupts after request_irq
- Restored MSI data register value at driver unload time
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Signed-off-by: Adrian Bunk <bunk@stusta.de>
|
|
Hi,
This patch contains additional statistics counters added to s2io driver
these statistics are very much usefull in debugging the driver.
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
Hi,
This patch contains some of the bug fixes done for S2io driver.
Following are the brief description of changes
1. Continuing with initialization if we get minimum required MSI-X vectors
2. fix for ethtool online link test fails
3. make wait_for_cmd_complete generic for all command status registers
4. Print "Device is on PCI-E bus" for Xframe-E card
5. CX4 requires additional delay after sw_reset, and requires higher value for igp
6. Fixed panic due to non-TCP and/or LLC/SNAP traffic in case of lro
7. remove legacy code for old transponder
8. SPECIAL_REG_WRITE made to use 32-bit writes irrespective of system type
9. handle link interrupt as per user guide for Xframe II
10. Wait till all interrupts hndled
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
hi,
This patch contains the modification and bug fixes with respect to
input parameters and outupt dmesages. following is brief description
of the changes.
1. Set default values for rx_ring_sz[0..7] and tx_fifo_len[0..7]
2. verify few basic load parameters
3. read product description from VPD
4. clean up of dmesg when driver is loaded
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
|
|
|
|
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: Ravinandan Arakali <ravinandan.arakali@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
This patch makes some needlessly global code static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
|
This patch implements the UFO support in S2io driver. This patch uses the UFO
interface available in linux-2.6.15 kernel.
Signed-off-by: Ananda Raju <ananda.raju@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
|
