File intel-e1000e.changes of Package intel-e1000e

File intel-e1000e.changes of Package intel-e1000e (Revision 0fb6cc9a6ddd14cdd733ea41781a4fa3)

Currently displaying revision 0fb6cc9a6ddd14cdd733ea41781a4fa3, show latest

-------------------------------------------------------------------
Sun Sep 23 18:21:55 UTC 2012 - cs@linux-administrator.com

- 2.1.4
  * 82577/8/9 - increase MSE threshold to avoid unexpected link drop
  * 82571 Tx Data Corruption during Tx hang recovery
  * 82571 serdes link instability if auto-negotiation fails - RH bug #771366
  * 82571 serdes link issue - RH #840642 and SF #346
  * Enable EEE on 82579/i217 only when link partner is capable of EEE
  * DoS while TSO enabled caused by link partner with small MSS
  * Program the correct register for ITR when using MSI-X on 82574
  * Test that 82577/82578/82579/i217 PHY is accessible
    otherwise driver can fail to load
  * remove use of the unnecessary IP payload checksum to allow concurrent
    use of receive hashing offload and jumbo frames
  * implement mdi-mdix setting supporton 82577/9
  * use more informative logging macros when netdev not yet registered
  * compile failure on 2.6.37+ kernels with CFLAGS_EXTRA="-DE1000E_NO_NAPI"
  * bug in bit shift when setting SMBus frequency on i217
  * Upstream commit 80be3129d7152fe73b7d5db4595e2f4267497f24
    - e1000e: add transmit timestamping support
  * Upstream commit 59aed95263bdd0e2b48eb9be5a94346d2d4abf90
    - e1000e: Remove special case for 82573/82574 ASPM L1 disablement
  * Upstream commit d4a4206ebbaf48b55803a7eb34e330530d83a889
    - e1000e: Disable ASPM L1 on 82574
  * Upstream commit 727c356f4d799b53f94cf8fe43e19d64482348c7
    - e1000e: Fix default interrupt throttle rate not set in NIC HW
  * Upstream commit 569a3aff70e880588fe4b3f1622ac60abbeb4a28
    - e1000e: MSI interrupt test failed, using legacy interrupt

-------------------------------------------------------------------
Tue Jun 19 09:29:13 UTC 2012 - jweberhofer@weberhofer.at

- 2.0.0.1
  * Add preliminary basic support for i217 available on some platforms
    with the Lynx Point Platform Controller Hub (PCH) chipset from Intel
  * Upstream commit b3f4d5990bfc8b060e5010c1464789fca1f4c5b4 - intel: make
    wired ethernet driver message level consistent 
  * fix and commonize 82577/8/9 and i217 PHY init/resume work
  * fix LPLU in non-D0a on 82577/8
  * cleanup code including new checkpatch warnings, boolean expressions and 
    whitespace issues
  * Upstream commit 4679026d783eb5ac90247bc466d66b817b213abf
    - net: remove the second argument of k[un]map_atomic()
  * Upstream commit cf955e6c96cb9af2dce7093a8e3a5e808555c8a4 
    - e1000e: Support RXALL feature flag.
  * Upstream commit 943146de22d3658de26d94512beae7223282ed1c
    - e1000e: Support sending custom Ethernet CRC
  * Upstream commit 0184039a4b6727d6efd545919c773ef141090ae7
    - e1000e: Support RXFCS feature flag
  * Upstream commit 3f0cfa3bc11e7f00c9994e0f469cbc0e7da7b00c
    - e1000e: Support for byte queue limits
  * Fix WARN_ON when resetting device while it is already being reset 
  * Fix 82577/8 autoneg'ing to 10Mbps instead of 1000Mbps in D0 
  * WoL fix for 82577/8 * Adjust delay after LANPHYPC toggle for 82579

-------------------------------------------------------------------
Sat Apr 21 07:12:18 UTC 2012 - jweberhofer@weberhofer.at

- 1.11.3
  * Enabled DMA Burst Mode on 82574 by default for performance gain with small packets.
  * Fixed issue with 82574/82583 sometimes not auto-negotiating gigabit link.
  * Disabled IPv6 extension header parsing because some malformed IPv6 headers can hang the Rx.
  * Disabled IBIST slave mode (far-end loopback) on 80003ES2LAN during reset to ensure the mode does not accidentally become enabled.

- 1.10.6
  * Fix kernel oops caused by concurrent reset and close of device (SourceForge bug tracker 3486658)
  * Code cleanup - whitespace, indentation, use true/false for booleans,
    use function pointers rather instead of API calls, checkpatch warnings,
    sparse warnings, remove unnecessary parentheses, duplicate code, use
    logical operators instead of comparing to true/false, remove unnecessary
    return from void functions, goto coding style
  * Upstream commit 5f4a780ddd453c4918555fed9d9c5f2d455a087d - e1000e: hitting BUG_ON() from napi_enable
  * Fix compilation issues when compiled with W=1
  * Upstream commit 9e2d7657e2a8fb40f732563dffb05151ea2d7e01 - e1000e: cleanup - check return values consistently
  * Upstream commit f23efdff77bbc1207255ccebe04771970604bbd9 - e1000e: cleanup e1000_set_phys_id
  * Upstream commit 0e15df490eef6f5080b84533dcd3068a78122768 - e1000e: minor whitespace and indentation cleanup
  * Upstream commit fe1e980f24697edb7d4e17cd74bbeae4a0388525 - e1000e: remove unnecessary returns from void functions
  * Upstream commit 75ce15327030128d713079d1f5e23cbe7af532f6 - e1000e: cleanup: rename goto labels to be more meaningful
  * Upstream commit 5015e53a4cf0c88977120faede7eb02b0459d90e - e1000e: cleanup goto statements to exit points without common work
  * Fix a possible device hang on 82579 during stress
  * Upstream commit c85c21ad1fe18daf26bfe8684df974d88cec50b7 - e1000e: Need to include vmalloc.h
  * Upstream commit 8e586137e6b63af1e881b328466ab5ffbe562510 - net: make vlan ndo_vlan_rx_[add/kill]_vid return error value
  * Upstream commit 09357b00255c233705b1cf6d76a8d147340545b8 - e1000e: Avoid wrong check on TX hang
  * WoL fixes for 82577/8/9
  * Fix for 82579 failure to read PHY register during init
  * Fix packet drop issue on 82579
-------------------------------------------------------------------
Wed Jan 25 08:58:30 UTC 2012 - jweberhofer@weberhofer.at

- 1.9.5
  * Fix compile issues on Fedora 15 kernel
  * Cleanup whitespace and other checkpatch warnings/errors
  * conditionally restart autoneg on 82577/8/9 when setting LPLU state 
  * Upstream: e1000e: Convert printks to pr_<level> (commit ef456f85)
  * Upstream: Sweep the last of the active .get_drvinfo floors under ethernet/
    (commit 612a94d6)
  * increase Rx PBA to prevent dropping received packets on 82566/82567
  * Upstream: net: introduce and use netdev_features_t for device features sets
    commit (c8f44aff)
  * ICHx/PCHx LOMs should use LPLU setting in NVM when going to Sx
  * update workaround for 82579 intermittently disabled during S0->Sx
  * Upstream: drivers/net: Add module.h to drivers who were implicitly using it
    (commit 9d9779e723a5d23b94abbe5bb7d1197921f6f3dd)
  * Fix build warnings with CFLAGS_EXTRA=-DE1000E_NO_NAPI
  * disable Early Receive DMA on ICH LOMs
  * Upstream: net: add skb frag size accessors (commit 9e903e085262ffbf1fc44a17ac06058aca03524a)
  * Upstream: intel: convert to SKB paged frag API (commit
    877749bf3f2f7a517ae74cd2c2fa4eed7aa9b51d)
  * Upstream: PM QoS: Move and rename the implementation files (commit
    e8db0be1245de16a6cc6365506abc392c3c212d4)
  * Upstream: e1000e: convert to stats64 (commit
    67fd4fcb78a7ced369a6bd8a131ec8c65ebd2bbb)
  * Upstream: net: Remove unnecessary driver assignments of ethtool_ringparam
    fields to zero (commit 8b0c11679fd37522d8d34a76101319a085d80912)
  * Upstream: e1000e: fix skb truesize underestimation (commit
    98a045d7e4a59db0865a59eea2140fe36bc7c220)
  * Upstream: e1000e: convert to real ndo_set_rx_mode (commit
    ef9b965a1c4ebd0f0ee961dbc328e1222dc8f487)
  * 82574/82583 Tx hang workaround
  * use hardware default values for Transmit Control register
  * Upstream: e1000e: locking bug introduced by commit 67fd4fcb (commit
    a90b412cb8c7ccc1689f9ea130883d00a1f0a5bb)
  * remove custom setting of TIPG for 8000es2lan
  * use default settings for Tx Inter Packet Gap timer
  * workaround WoL failures on some ICH8 LOMs
  * 82579: workaround for link drop issue
  * always set transmit descriptor control registers the same
  * re-factor ethtool get/set ring parameter
  * pass pointer to ring struct instead of adapter struct
  * default IntMode based on kernel config & available hardware support
  * convert head,	tail and itr_register offsets to __iomem pointers
  * remove MSI-X code not wrapped with CONFIG_E1000E_SEPARATE_TX_HANDLER to sync
    with upstream driver
  * re-enable alternate MAC address for all devices which support it
  * Upstream: e1000e: convert to netdev features/hw_features API (commit
    dc221294719ae0f28cc260cc37edd439161088a9)
  * cleanup Rx checksum offload code
  * Upstream: net: remove use of ndo_set_multicast_list in drivers (commit afc4b13df143122f99a0eb10bfefb216c2806de0)
  * workaround Rx crc errors with jumbo frame stress on 82578
  * add Receive Packet Steering (RPS) support

-------------------------------------------------------------------
Thu Oct 27 20:24:24 UTC 2011 - jweberhofer@weberhofer.at

- 1.6.3
  * Fix for rx_no_buffer_count increments when RPS is enabled

- 1.6.2
  * 82578DM: Fix for WOL fails after several reboots. 
  * Enable Rx hash use by stack for 1Gig.
  * Convert to new VLAN model.
  * 82577: rx_crc_errors wirh jumbo frames and SSC enabled.

- 1.5.1
  * Do not support Alternate MAC Address if NVM word 0x37 has a value of 0x0000
  * Change workaround for Tx hang in previous release to do same for Rx side
  * Two workarounds for missed packets on 82579 in 100half
  * Perform write flush after most registers writes that are followed by a delay
  * Fix for panic on RHEL5.7 with vlan code

- 1.4.4
  * 82579: Fix for Tx Hang on FTS ME Platform
  * 82571: Add device id to fix for alternate MAC address feature
  * Upstream commits: 25db0338, 70739497 and 14ad2513, Change ethtool interfaces to use ethtool_cmd_speed[_set] API.
  * Convert use of receive descriptors from legacy to extended format
    to allow future enablement of advanced features such as timesync and RPS.

- 1.3.17
  * Fixed an occassional panic on RHEL5.4 and newer when suspending/hibernating the adapter.
  * 82577/8 clear Host Wakeup bit without touching PHY registers on page 800 (necessary as a workaround)
  * Upstream: PM: Remove CONFIG_PM_OPS (commit aa338601)
  * Upstream: e1000e: If ASPM L0s needs to be disabled, do it prior to enabling device (commit 78cd29d5)
  * Use new ethtool set_phys_id entry point
  * Do not set blink_led function pointer incorrectly for 82574/82583
  * Add ability to access multiple PHY registers on the same page while holding a lock without having to set the page for every access. Use this method in a number of places for ICH/PCH adapters. 
  * 82579: Fixed an issue where the adapter could become disabled during Sx stress testing. Renamed e1000_disable_gig_wol() to better reflect what it does.
  * Upstream: e1000e: use dev_kfree_skb_irq() instead of dev_kfree_skb() (commit ef5ab89c)
  * Upstream: net: use pci_dev->revision, again (commit ff938e43)
  * Convert msec_delay's less than 20msec to usleep_range as described in ./Documentation/timers/timers-howto.txt in the kernel source tree
  * Refactor fix in previous release that fixed an issue where the adapter could reset unnecessarily when ethtool -r is done while passing traffic
  * Report PCIe link speed during driver load in GT/s, not GB/s
  * Do not report driver supports Wake-on-ARP since it does not
  * Various code cleanups
  * ICH/PCH: release mutex but not sw/fw/hw semaphore (swflag) after hw reset; log debug message if swflag is released unexpectedly
  * 80003es2lan: fixed an issue where the ESB2 LAN interface could become non- functional and possibly identify it's MAC address incorectly to switches when setting up the switching tables

-------------------------------------------------------------------
Mon Feb 14 15:06:00 CEST 2011 - opendevel@weberhofer.at

- 1.3.10a

- 82579: Don't allow jumbo frames to be configured when MACsec enabled 
in EEPROM

- 82567LM-3/82567LF-3/82567V-4: use MAC-based WoL instead of PHY-based 
WoL. The unsupported PHY-based WoL was inadvertently enabled when 
support was added for 82577/82578, and could cause problems in Sx 
states.

- Do not increment tx_timeout statistic when normally resetting the 
adapter when there is outstanding transmit work.

- Do not run any scheduled tasks when adapter is going down, otherwise a 
deadlock may occur.

- Miscellaneous comment/printk typos

- 82579: Update EEE timer value to prevent link from dropping

- Various minor changes based on static analysis of the driver code

- Prevent adapter from being reset when passing traffic and re-starting 
auto-negotiation via 'ethtool -r ethX'

- Upstream: don't use flush_scheduled_work (commit 23f333a2)

- Upstream: Fix drivers advertising HW_CSUM feature to use csum_start 
(commit 0d0b1672)

- Upstream: fix double initialization in blink path (commit 9633e63b)

- Upstream: Use default implementation of ethtool_ops::get_link where 
possible (commit ed4ba4b5)

- Upstream: add netpoll support for MSI/MSI-X IRQ modes (commit 
147b2c8c)

- Use kmemdup rather than duplicating its implementation

- Fix sparse warnings

- 82579: fails ethtool online diagnostic link test when connected to 
certain switches. Extend the timeout to acquire link to 5 seconds and 
document the issue.

- Use ether_crc_le rather than duplicating its implementation

- 8257x: do not ignore return from e1000_get_phy_id)82571()

- 82571: PHY_CONTROL bit 11 (power down) cleared when hardware is reset. 
Power down the PHY during reset when the interface is down, not running 
ethtool diagnostics and WoL is disabled.

- Use vzalloc instead of vmalloc/memset

- Rename files to match closer to upstream driver filenames

- Fix non-whitespace checkpatch issues

- Fix compile warnings with new gcc 4.6

- Declare various tables as static const to reduce data and overall 
object size

- 82574: requires its own set_d[0|3]_lplu_state functions

- Fix compilation errors on SuSE SLES10 PPC

- Fix spec file to allow rpm builds on newer OSV

- Upstream: avoid some skb->ip_summed initializations (commit bc8acf2c)

- Upstream: return operator cleanup (commit 807540ba)

- Upstream: set NETIF_F_HIGHDMA for VLAN feature flags (commit 7b872a55)

- Upstream: Remove address use from assignments of function pointers 
(commit c061b18d)

- Upstream: Don't check for vlan group before vlan_tx_tag_present 
(commit eab6d18d)

- Upstream: Rename VLAN_GROUP_ARRAY_LEN to VLAN_N_VID (commit b738127d)

- Upstream: Simplify MSI interrupt testing (commit 068e8a30)

- 82574: do not disable ASPM L1

- 82571/82572/80003ES2LAN: use hardware writeback batching to improve 
bus utilization and removes partial writes on certain systems.

-------------------------------------------------------------------
Thu Nov 25 23:06:00 CEST 2010 - opendevel@weberhofer.at

- 1.2.20 Version upgrade

-------------------------------------------------------------------
Tue Apr 20 17:38:56 CEST 2010 - linux@weberhofer.at

- update to build 1.1.2.1a

-------------------------------------------------------------------
Tue Sep  8 16:16:56 CEST 2009 - sbahling@suse.de

- initial build of 1.0.2.5