Commit graph

33 commits

Author SHA1 Message Date
Benjamin LaHaise
cc104e52b4 vxge: Version update
Update the driver version number for any bug reports from end users.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:46 -07:00
Benjamin LaHaise
ea11bbe0c9 vxge: prefetch skb->data
This patch implements prefetching of skb->data from a copy of the pointer
in the descriptor (which is already in the L1 cache at this point).  This
improves netperf rx performance (netperf -L 0,0 -c -H 192.168.254.2 -- -M
131072 -m 131072) by 4.9% on a P4 Xeon host.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:45 -07:00
Benjamin LaHaise
3f23e436d2 vxge: prefetch RxD descriptors
This patch prefetches RxD descriptors which helps to lower the latency of a
cache miss in vxge_hw_ring_rxd_next_completed.  This lowers the % of CPU
time used by vxge_hw_ring_rxd_next_completed() where the descriptor is
accessed in profiling netperf on a P4 Xeon from 1.5% to 1.0%.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:45 -07:00
Benjamin LaHaise
ff1b974cf3 vxge: improve tx performance by using mmiowb() instead of wmb()
wmb() is extremely heavy on x86.  The semantics required in the driver are
provided by mmiowb(), so use that and improve tx performance on P4 Xeons by
5-10%.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:44 -07:00
Benjamin LaHaise
ff67df55f6 vxge: fix pktgen hangs (don't abuse skb->cb[])
This patch fixes a case in the transmit completion code which was resulting
in pktgen hanging at the end of a run.  The cause is due to the fact that
the ->cb[] area of an skb cannot be used in a network driver's transmit
path, as that area belongs to the network protocol.  Pktgen hangs, as it
sends out the same packet multiple times, and vxge's use of this area of
the skb for a temporary list can only add the packet to the temporary list
once (while it may be on the queue many times).  The fix is to remove this
abuse of skb->cb[].  Instead, skb pointers are placed into a temporary
stack array, and then free outside of the tx lock.  This retains the smp
optimization of doing dev_kfree_skb() outside of the tx lock.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:43 -07:00
Benjamin LaHaise
4403b37188 vxge: don't drop frame on tx queue full
The vxge driver will drop a packet in its transmit function if the number
of TxDs available hits 0.  Instead of doing that, simply stop the transmit
queue when transmitting a packet with the last available TxD.

Signed-off-by: Benjamin LaHaise <ben.lahaise@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-05 13:10:43 -07:00
Dean Nelson
e33b992d15 vxge: return PCI_ERS_RESULT_DISCONNECT on permanent failure
PCI drivers that implement the struct pci_error_handlers' error_detected
callback should return PCI_ERS_RESULT_DISCONNECT if the state passed in is
pci_channel_io_perm_failure. This patch fixes the issue for vxge.

Signed-off-by: Dean Nelson <dnelson@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-01 23:12:12 -07:00
Sreenivasa Honnur
a73e76e23c vxge: Version update
- Version update

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:44 -07:00
Sreenivasa Honnur
0a25bdc696 vxge: Printing the function's configured mode of operation
- Printing the function's configured mode of operation.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:41 -07:00
Sreenivasa Honnur
22fa125ee2 vxge: Disable fcs stripping.
- Disable fcs stripping. The minimum frame size that can be received by the
hardware is 57 Bytes. A 64 Byte Ethernet frame with the vlan tag and fcs
stripped will result in a 56 Byte frame which will lock up the receive
engine. The work around is to disable fcs stripping in the hardware which
is done with a firmware upgrade. The fixes are -
1. Ensure that the correct firmware version is used.
2. Decrement the indicated packet length of the receive packet by 4 bytes
(FCS length).

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:38 -07:00
Sreenivasa Honnur
a5d165b571 vxge: Fixes in isr routine
- Fixes in isr routine
   Fixed crash with INTA mode during driver unload. Pass hldev to request_irq
   instead of vdev.
   Fixed Traffic failure with GRO in INTA mode. Pass the correct napi
   context to gro functions.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:34 -07:00
Sreenivasa Honnur
6052ae1676 vxge: Removed the code to bounds check the mss value
- Removed the code to bounds check the mss value. The hardware does bounds checking and
will not allow an oversized mss to lockup the transmit path.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:32 -07:00
Sreenivasa Honnur
3255da416d vxge: Corrected Register map entry for xmac_link registers
- Corrected Register map entry for xmac_link registers.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:30 -07:00
Sreenivasa Honnur
bd9ee6808b vxge: check for card status before continuing in device close
- Fixed the crash in rmmod after vpath open failed when trying to change mtu.
We should check for card status before continuing in device close.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:28 -07:00
Sreenivasa Honnur
914d0d714f vxge: Fixed memory leak by freeing memory allocated for device_config
- Fixed memory leak by freeing memory allocated for device_config.

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:26 -07:00
Sreenivasa Honnur
7975d1eed7 vxge: Removed ioremap of unused bar addresses and their references
- Removed ioremap of bar1 address
  Driver needs only bar0 address for register access
- Removed references to bar1 and bar2 addresses

Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:24 -07:00
Patrick McHardy
6ed106549d net: use NETDEV_TX_OK instead of 0 in ndo_start_xmit() functions
This patch is the result of an automatic spatch transformation to convert
all ndo_start_xmit() return values of 0 to NETDEV_TX_OK.

Some occurences are missed by the automatic conversion, those will be
handled in a seperate patch.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-07-05 19:16:04 -07:00
Sivakumar Subramani
5dbc901172 vxge: Enable SRIOV support in the driver.
- Enabled SRIOV support in the driver.
- Call __vxge_hw_verify_pci_e_info() for the PF only. This function
  verifies the negotiated link width and current link speed in the
  Link Status Register (offset 12h) which are reserved fields for VFs
  as per the SRIOV specification, section 3.5.8.
- Implemented David Miller's comment to remove the #ifdef CONFIG_PCI_IOV as
  these intefaces have NOP versions declared when the defintion is not set.

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-06-17 04:31:14 -07:00
Figo.zhang
50d36a93dc vxge: No need for check vfree() pointer.
vfree() does its own 'NULL' check, so no need for check before
calling it.

Signed-off-by: Figo.zhang <figo1802@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-06-11 02:47:14 -07:00
Eric Dumazet
28679751a9 net: dont update dev->trans_start in 10GB drivers
Followup of commits 9d21493b4b
and 08baf56108
(net: tx scalability works : trans_start)
(net: txq_trans_update() helper)

Now that core network takes care of trans_start updates, dont do it
in drivers themselves, if possible. Multi queue drivers can
avoid one cache miss (on dev->trans_start) in their start_xmit()
handler.

Exceptions are NETIF_F_LLTX drivers (vxge & tehuti)

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-05-29 01:46:26 -07:00
David S. Miller
bb803cfbec Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Conflicts:
	drivers/scsi/fcoe/fcoe.c
2009-05-18 21:08:20 -07:00
roel kluin
d77dd8d27e Neterion: *FIFO1_DMA_ERR set twice, should 2nd be *FIFO2_DMA_ERR?
FIFO1_DMA_ERR is set twice, the second should be FIFO2_DMA_ERR.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Acked-by: Ram Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-05-17 21:18:05 -07:00
David S. Miller
a4fe91ee71 Revert "vxge: use max() instead of VXGE_HW_SET_LEVEL"
This reverts commit 011983048a.

Causes warnings in the build as reported by Stephen Rothwell.
So this change is worse than what it's curing.

Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-29 17:53:20 -07:00
Alexander Beregalov
011983048a vxge: use max() instead of VXGE_HW_SET_LEVEL
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: Ram Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 03:02:47 -07:00
Alexander Beregalov
4e204c1000 vxge: fix possible NULL dereference in vxge-traffic.c
If vpath is NULL then hldev is NULL also.

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: Ram Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 03:02:46 -07:00
Alexander Beregalov
2b05e0021b vxge: should include tcp.h
Fix this build error:

drivers/net/vxge/vxge-main.c: In function 'vxge_get_vpath_no':
drivers/net/vxge/vxge-main.c:695: error: dereferencing pointer to
incomplete type
...

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-04 16:36:18 -07:00
Ramkrishna Vepa
f1782852ad Neterion: New driver: Kconfig and Makefile
- Kconfig and Makefile related changes for vxge driver.

- No changes in current submission.

- Changes from previous submission -
 - Incorporated the following review comments as per Bill Flink:
        - Add dependancy on INET along with PCI
        - Remove dependancy on INET_LRO and add GRO support.
 - Made this patch as last patch as per Ben Hutchings comments.

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:46 -07:00
Ramkrishna Vepa
5f548a3281 Neterion: New driver: Driver version
- Driver version

- Changes from previous submission -
 - Updated driver version.

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:45 -07:00
Ramkrishna Vepa
d26078fe72 Neterion: New driver: Ethtool related
This patch implements all ethtool related entry point functions for the driver.

- Changes in this submission -
        - Removed redundant #include statements

- Changes in previous submissions -
        - Removed unused statistics.
        - General clean up - removed unused variables and macros.
 - Incorporated following comments from Ben Hutchings
        - No need to restart the interface in vxge_ethtool_sset
        - Do not use #ifdef ADVERTISED_XX
        - Remove unnecessart intermediate copy in vxge_ethtool_gdrvinfo
        - Use strlcpy() to ensure null-termination.
        - Use ethtool_op_get_tso, ethtool_op_set_tx_hw_csum instead
          of redefining the same code.
        - Implement get_strings_count instead of deprecated get_stats_count.
        - Don't report reporting the EEPROM length as we don't supprt
          get/set eeprom.
        - Don't set self_test_count as we don't support any self tests.

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:44 -07:00
Ramkrishna Vepa
703da5a1a2 Neterion: New driver: Main entry points
This patch implements all the driver entry point functions.
- Definition of all module loadable paramters.
- Implementation of all driver entry point functions.

- Changes in this submission -
        - Fixed compilation error when enabling debug statements
        - Fixed few warnings when CONFIG_PCI_MSI is not defined
        - Removed unnecessary volatile variables
        - Removed compare_ether_addr as it causes unaligned memory access on
          the sparc64 platform.

- Changes in previous submissions -
        - As per Stephen Hemminger's comments removed the following loadable
          parameters - gro, rx_& tx max_indicate_pkts, exec_mode, rx & tx
          pause_enable, tx_steering_type and intr_type.
        - Added Device ID definition in vxge-main.h instead of
          include/linux/pci_ids.h file - Reported by David Miller
 - Incorporated following review comments from Ben Hutchings
        - NAPI is always enabled (no option to turn it OFF).
        - Loadable parameters
         rx_steering_type: This loadable option is removed.
         ring_blocks: This loadable option is removed.
         The driver default settings work well in most if not all cases.
         Another patch to configure these parameters with ethtool will be
         released in the future.
 - LRO has been deprecated in favour of GRO - Bill Fink & Dave Miller's comment
 - Fixed sparse warnings - Reported by Andi Kleen
 - Removed unused variables

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:44 -07:00
Ramkrishna Vepa
113241321d Neterion: New driver: Traffic & alarm handler
This patch takes care of trafic handling related APIS.
- Interrupt Enable and disable
- Mask / Unmask Interrupt
- Traffic Interrupt handling.
- Alarm Interrupt handling.

- Changes in this submission -
        - General clean up - removed redundant includes, defines and macros.

- Changes in previous submissions -
 - General cleanup - removed unused functions and variables.
 - Use asserts where necessary - Reported by Andi Kleen
 - Fixed sparse warnings - Reported by Andi Kleen
 - Use a prefix, "__vxge" in front of hw functions to make them globally
   unique - Ben Hutchings

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:43 -07:00
Ramkrishna Vepa
40a3a9156d Neterion: New driver: Hardware init & configuration
This patch takes care of Initialization and configuration steps of
Neterion Inc's X3100 Series 10GbE PCIe I/O Virtualized Server Adapter.
- Device Initialization.
- Verification and setting of device config parameters.
- Allocation of Tx FIFO and Rx Ring descriptors (DTR).
- APIs to get various type of hw stats
- APIs to configure RTS (Receive Traffic Steering)

- Changes in this submission -
        - Include vmalloc header without which a compilation error occured
          on sparc64, ppc64 and IA64 plaforms.
	- Fixed compilation warning in register_poll, write32_upper,
          write32_lower and the special write64 functions on ppc64.
        - General cleanup - removed redundant includes and defines.

- Changes in previous submissions -
        - Add readq/writeq implementation for the driver for 32 bit systems -
          reported by Dave Miller.
- Incorporated following comments from Ben Hutchings
        - Start a comment with "/**" to make it a kernel-doc comment.
        - Use prefix, "__vxge" in front of hw functions to make them globally
          unique.
        - Fixed unnecessary clearing members of *channel just before freeing
        - Use backslashes only for macro definitions and not in multi-line
          statements.
        - Used pci_find_capability instead of redefining it.
        - Used device and revision ids that are already in pdev - no need to
          read them again.
        - Used pci_save_state() and pci_restore_state() around resets.
        - Used udelay and mdelay directly instead of wrapper.
        - In __vxge_hw_device_register_poll() reset i to 0 after the
          microsecond delay loop to commence the millisecond delay loop.
        - Corrected spelling "sapper" - should be "swapper"
        - Remove too much vertical whitespace.
        - Replaced magic numbers with appropriate macros
- Incorporated following comments from Andi Kleen [andi@firstfloor.org]
        - Reduced the arguments in functions or refactored them into smaller
        functions.
        - Allocate page sized memories used in slow path with vmalloc.
        - Use asserts where necessary.
        - Use macros instead of magic numbers.
        - Use the pci layer code instead of defining own functions
        - Remove driver wrappers such as xge_hw_device_private_set().
        - Fixed sparse warnings.

Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:41 -07:00
Ramkrishna Vepa
66d97fedea Neterion: New driver: register set - vxge-reg.h
- Complete Register map details of Neterion Inc's X3100 Series 10GbE PCIe I/O
  Virtualized Server Adapter.

- No change from previous submission.

- Changes in previous submissions -
- Incorporated following comments from Ben Hutchings
        - Use original macros for endian checks
        - Remove VXGE_OS_PLATFORM_* macros as they are unused.
        - Converted multiple bVALX macros into single with additional
          width parameter and renamed it to vxge_bVALn.
        - Using __packed instead of pragma pack(1)
        - Added a comment of the use of a hw swapper so driver code is
          portable (does not have to change the byte order for register
          access as well as dma operations) on different ENDIAN platforms.
        - Using the <linux/pci_regs.h> definitions instead of redefing them.
        - Using the PCI capabilities registers in <linux/pci_regs.h>
          instead of redefing them.
Signed-off-by: Sivakumar Subramani <sivakumar.subramani@neterion.com>
Signed-off-by: Rastapur Santosh <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-02 00:33:41 -07:00