Search
j0ke.net Open Build Service
>
Projects
>
hardware
:
nic
>
intel-e1000e
> intel-e1000e.changes
Sign Up
|
Log In
Username
Password
Cancel
Overview
Repositories
Revisions
Requests
Users
Advanced
Attributes
Meta
File intel-e1000e.changes of Package intel-e1000e
------------------------------------------------------------------- Fri Feb 7 14:40:38 UTC 2014 - cs@linux-administrator.com - 2.5.4 * Add - ether_addr_equal for use by versions less than 3.5 * Fix - Semaphore imbalance for 82573 devices * Fix - 82579 ethtool offline register tests reg bitmasks * Fix - Perform PHY configuration after exiting ULP mode on I218 * Cleanup - braces {} should be used on all statement arms or none * Fix - high CRC error count with 4K RX traffic on 82577/8 * Fix - overrun of PHY RAR array on 82579 * Add - Adding error reporting to init_phy_workarounds_pchlan() * Fix - Defaulting to continue on ME blocked reset * Fix - Hardware Unit Hang * Compat - PCI_EXP_LNKCTL_ASPMC introduced in kernel 2.6.29 * Fix - Use macro instead of digit for defining e1000_rx_desc_packet_split * Fix - Remove call for setting MDIO slow mode for I217/I218 * Fix - I218 ULP updates * Fix - Incorrect entries in system log on driver load with cable out * Fix - Explicitly set Transmit Control Register to ensure 82574L 100M HDX performance * Fix - 82579 sets LPI too early * Fix - vlan support for kernels 3.10 and above * Fix - for new checkpatch JIFFIES_COMPARISON warning * Add - additional I217/I218 PCI IDs * Fix - Using InterruptThrottleRate=2 no error message and ITR not set * Add - missing branding strings in ich8lan.c * Fix - ULP is not supported on LPT MAC with I218 PHY * Fix - build warnings on SLES11 SP3 - 2.4.14 * Add support for I218 "Ultra-Low Power" (ULP) mode enabled in S3/4 and in S0 when Runtime Power Management (RPM) is enabled * Fix - do not access MAC registers when RPM suspended * Fix - workaround a Tx hang inssue at 100Half on I218 * Fix - disable ASPM L1 on 82583 * Cleanup - remove unnecessary workaround for 82578 * Cleanup/refactor - Runtime Power Management flow * Fix - device would repeatedly cycle D0->D3 when RPM enabled and the interface was administratively downed * Fix - do not wake from RPM to read PHY status * Fix - release mutex lock only if it has been initially acquired on 82574/82583 * Upstream - commit 73e3dd6b45c4c870fc2641eb04c24e3f12dab1e0 (e1000e: fix numeric overflow in phc settime method) * Upstream - commit 80d5c3689b886308247da295a228a54df49a44f6 (net: vlan: prepare for 802.1ad VLAN filtering offload) * Upstream - commit f646968f8f7c624587de729115d802372b9063dd (net: vlan: rename NETIF_F_HW_VLAN_ feature flags to NETIF_F_HW_VLAN_CTAG_) * Upstream - commit 86a9bad3ab6b6f858fd4443b48738cabbb6d094c (net: vlan: add protocol argument to packet tagging functions) * Fix - AMT connection drops after driver unload if large MTU configured * Upstream - commit 37287fae7e4506fb0822523587c769783164cdd6 (e1000e: Add missing dma_mapping_error-call in e1000_alloc_jumbo_rx_buffers) * Fix - low throughput using 4K jumbo MTU * Fix - error checking in initialization flow for I218 * Fix - ethtool interrogating device takes it out of RPM D3 briefly * Refactor/fix system hibernate flow * Fix - automatically enable EEE LPI in I217/I218 hardware when in Sx * Cleanup - whitespace, indentation and lines exceeding 80 characters * Fix - Resolve strict checkpatch USLEEP_RANGE checks * Upstream - commit e60b22c5b7e59db09a7c9490b1e132c7e49ae904 (e1000e: fix accessing to suspended device) * Upstream - commit 66148babe728f3e00e13c56f6b0ecf325abd80da (e1000e: fix runtime power management transitions) * Upstream - commit 4e0855dff094b0d56d6b5b271e0ce7851cc1e063 (e1000e: fix pci-device enable-counter balance) * Fix - do not enable EEE on 82579/I217/I218 for speeds in which EEE is not advertised * Fix - IEEE MMD register 3.1 at wrong EMI address on 82579 ------------------------------------------------------------------- Fri May 10 14:50:36 UTC 2013 - cs@linux-administrator.com - 2.3.2 * EEE status inaccurately reported by ethtool when connected back-to-back with another e1000e-supported device with EEE * Fix compilation issues on SLES11 SP3 * Upstream commit ffe0b2ff17e1d59d33334767e07e8141de4bd5c8 - e1000e: Use standard #defines for PCIe Capability ASPM fields * Upstream commit e792cd916cf74315bddb2c8b2323ef498cd7bfde - e1000e: display a warning message when SmartSpeed works * Fix compilation issues on RHEL6.4 * Cleanup 'unused parameter' compilation warnings * Fix DMA unit hang on I218 * Use kernel-defined MII bit/register definitions * Fix performance issue between two 82579 systems connected via a 10Mbps hub * Fix link setup code for I217/I218 * Fix file permissions * Upstream commit aaeb6cdfa5c07533c2cd6d2c381374c69f7db9dc - remove init of dev->perm_addr in drivers * Upstream commit 9f9a12f8ca79839c948464a37c5b557808278708 - net/intel: remove __dev* attributes * Fix LED blink on designs with LEDs driven by cathode * Code cleanups ------------------------------------------------------------------- Sun Feb 10 10:52:35 UTC 2013 - cs@linux-administrator.com - 2.2.14 * enable ECC on i217/i218 and reset the adapter on unrecoverable ECC error * checkpatch cleanups * change multi-line comment blocks to conform to preferred style in the in-kernel networking code * Upstream commit 6e97c170db3878695cb48b5021fffe442ec0aeb0 - e1000e: Minimum packet size must be 17 bytes * Upstream commit 3646f0e5c97703cecfd96cdabb085e78d9b7f680 - netdev: make pci_error_handlers const * Enable timesync (IEEE 1588 PTP) support on 82574/82583/82579/i217/i218 * remove redundant ethtool statistic counter (rx_long_byte_count, use existing rx_bytes instead) * Fix writes of large blocks to EEPROM on 80003es2lan/82571/82572 * Fix compilation errors on 3.7 and newer kernels * Upstream commit 8edc0e624db3756783233e464879eb2e3b904c13 - e1000e: Change wthresh to 1 to avoid possible Tx stalls * Fix adapter resets while transmitting UDP traffic and link drops * Performance improvements for jumbo frames on 82566/82567/82577/82578/82579 * Add support for i218 * Fix bug determining kernel version when compiling * Add ethtool support for EEE * Allow EEE in MDI-X mode on 82579/i217/i218 * Fix bug in ethtool offline register test on i217/i218 * Fix 82571 serdes link issue * Upstream commit 4e8186b68fb944ad9e7fd4080cd8bd8f10eb7cbd - e1000e: implement MDI/MDI-X control * Upstream commit f0c5dadff3fbda77a65b8577fee437c3d771233d - e1000e: fix panic while dumping packets on Tx hang with IOMMU * Upstream commit 49ce9c2cda18f62b13055dc715e7b514157c2da8 - drivers/net/ethernet: Fix (nearly-)kernel-doc comments for various functions * Upstream commit 7b1115e017569219eccbcad2bc48b81e6b98df29 - e1000e: advertise transmit time stamping ------------------------------------------------------------------- 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