- 14 Dec, 2015 16 commits
-
-
Rasmus Villemoes authored
0-termination is redundant, since sprintf has done that. Signed-off-by:
Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Ursula Braun authored
When sending skbs, qeth determines the number of qdio SBAL elements required. If qeth sends a fragmented skb, the SBAL element number calculation is wrong, because the fragmented data part is added twice in qeth_l3_tso_elements(). This patch makes sure fragmented data is handled in qeth_elements_for_frags() only, while qeth_l3_tso_elements() starts calculation of qdio SBAL elements just with the linear data part of the skb. Signed-off-by:
Thomas Richter <tmricht@linux.vnet.ibm.com> Signed-off-by:
-
Ursula Braun authored
/sys/class/net/<interface>/operstate for an active qeth network interface offen shows "unknown", which translates to "state UNKNOWN in output of "ip link show". It is caused by a missing initialization of the __LINK_STATE_NOCARRIER bit in the net_device state field. This patch adds a netif_carrier_off() invocation when creating the net_device for a qeth device. Signed-off-by:
Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reference-ID: Bugzilla 133209 Signed-off-by:
-
Thomas Richter authored
There have been 2 identical versions of function qeth_get_setassparms_cmd() for layer 2 and layer 3. Remove the layer 3 function qeth_l3_get_setassparms_cmd() and call the common one named qeth_get_setassparms_cmd() located in qeth_core_main.c Signed-off-by:
Eugene Crosser <Eugene.Crosser@ru.ibm.com> Signed-off-by:
-
Markus Elfring authored
The channel_remove() function tests whether its argument is NULL and then returns immediately. Thus the test around the calls is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
Markus Elfring <elfring@users.sourceforge.net> Signed-off-by:
-
Ursula Braun authored
Signed-off-by:
-
Eugene Crosser authored
When the linear buffer of the received sk_buff is shorter than the header, use skb_linearize(). sk_buffs with short linear buffer happen on the sending side under high traffic, and some kernel configurations, when allocated buffer starts just before page boundary, and IUCV transport has to send it as two separate QDIO buffer elements, with fist element shorter than the header. Signed-off-by:
-
Eugene Crosser authored
Initialize storage for the future IUCV header that will be included in the transmitted packet. Some of the header fields are unused with HiperSockets transport, and will contain data left from some other functions. Signed-off-by:
-
Alexander Aring authored
This patch removes ARPHRD_IEEE802154 from addrconf handling. In the earlier days of 802.15.4 6LoWPAN, the interface type was ARPHRD_IEEE802154 which introduced several issues, because 802.15.4 interfaces used the same type. Since commit 965e613d ("ieee802154: 6lowpan: fix ARPHRD to ARPHRD_6LOWPAN") we use ARPHRD_6LOWPAN for 6LoWPAN interfaces. This patch will remove ARPHRD_IEEE802154 which is currently deadcode, because ARPHRD_IEEE802154 doesn't reach the minimum 1280 MTU of IPv6. Also we use 6LoWPAN EUI64 specific defines instead using link-layer constanst from 802.15.4 link-layer header. Cc: David S. Miller <davem@davemloft.net> Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru> Cc: James Morris <jmorris@namei.org> Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org> Cc: Patrick McHardy <kaber@trash.net> Signed-off-by:
Alexander Aring <alex.aring@gmail.com> Signed-off-by:
-
Kazuya Mizuguchi authored
AVB-DMAC Receive FIFO Warning interrupt is not enabled, so it is not necessary to disable the interrupt in ravb_close(). On the other hand, this patch disables the interrupt in ravb_dmac_init() to prevent the possibility that the interrupt is issued by the state that a boot loader left. Signed-off-by:
Kazuya Mizuguchi <kazuya.mizuguchi.ks@renesas.com> Signed-off-by:
Yoshihiro Kaneko <ykaneko0929@gmail.com> Acked-by:
Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by:
-
Tejun Heo authored
bd1060a1 ("sock, cgroup: add sock->sk_cgroup") added global spinlock cgroup_sk_update_lock but erroneously skipped initializer leading to uninitialized spinlock warning. Fix it by using DEFINE_SPINLOCK(). Signed-off-by:
Tejun Heo <tj@kernel.org> Reported-by:
Dexuan Cui <decui@microsoft.com> Fixes: bd1060a1 ("sock, cgroup: add sock->sk_cgroup") Signed-off-by:
-
Fabian Frederick authored
Some eisa_driver structures used __init probe functions which generates a warning and could crash if function is called after being deleted. Signed-off-by:
Fabian Frederick <fabf@skynet.be> Signed-off-by:
-
Haiyang Zhang authored
In commit 2a04ae8a ("hv_netvsc: remove locking in netvsc_send()"), the locking for MSD (Multi-Send Data) field was removed. This could cause a race condition between RNDIS control messages and data packets processing, because these two types of traffic are not synchronized. This patch fixes this issue by sending control messages out directly without reading MSD field. Signed-off-by:
Haiyang Zhang <haiyangz@microsoft.com> Reviewed-by:
K. Y. Srinivasan <kys@microsoft.com> Signed-off-by:
-
Tom Herbert authored
Add support to enable and disable UDP checksums via netlink. This is similar to how VXLAN and GUE allow this. This includes support for enabling the UDP zero checksum (for both TX and RX). Signed-off-by:
Tom Herbert <tom@herbertland.com> Signed-off-by:
-
Tom Herbert authored
Fixe a couple of cast errors found by sparse. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 100GbE Intel Wired LAN Driver Updates 2015-12-13 This series contains updates to fm10k only. Jacob updates the driver to use ether_addr_copy() instead of copying byte-by-byte in a for loop. Fixed up CamelCase variable names and coding style issues. Cleaned up namespace pollution of fm10k_iov_msg_data_pf(). Cleaned up, by making it consistent, the use of VLAN and VLAN ID instead of vlan or vid. Lastly, update the driver to initialize XPS so that we can take advantage of the kernel feature. Alex Duyck fixed up the driver to free the resources associated with the MSI-X vector table if the q_vector allocation failed. Then fixed the driver to check for msix_entries to be NULL and freed the IRQ if the mailbox API returned an error on trying to connect. Bruce cleans up whitespace and namespace pollution issues in the driver. Also updates the driver to use the BIT() macro instead of bit-shifting coding. ==================== Signed-off-by:
-
- 13 Dec, 2015 24 commits
-
-
Jacob Keller authored
Similar to ixgbe and i40e, initialize XPS on driver load so that we can take advantage of this kernel feature. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Reviewed-by:
Bruce Allan <bruce.w.allan@intel.com> Tested-by:
Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Bruce Allan authored
Signed-off-by:
-
Bruce Allan authored
Make functions that should be static. While we're at it, fix the function header comment for fm10k_tlv_attr_nest_stop(), and update the copyright header for fm10k_pf.h, fm10k_tlv.c and fm10k_tlv.h. Signed-off-by:
-
Bruce Allan authored
Signed-off-by:
-
Bruce Allan authored
Signed-off-by:
-
Bruce Allan authored
The function declaration does not need to be 'inline'd here. Signed-off-by:
-
Alexander Duyck authored
This patch addresses two issues. First is the fact that the fm10k_mbx_free_irq was assuming msix_entries was valid and that will not always be the case. As such we need to add a check for if it is NULL. Second is the fact that we weren't freeing the IRQ if the mailbox API returned an error on trying to connect. Signed-off-by:
Alexander Duyck <aduyck@mirantis.com> Reviewed-by:
-
Alexander Duyck authored
If the q_vector allocation fails we should free the resources associated with the MSI-X vector table. Signed-off-by:
-
Jacob Keller authored
Rather than wrapping fm10k_dcbnl.c and fm10k_debugfs.c support with #ifdef blocks, just conditionally include the .o files in the Makefile. Also, since we're modifying it, update the copyright year on the Makefile. Signed-off-by:
-
Jacob Keller authored
We haven't bumped the driver version in a while despite many fixes being pulled in from the out-of-tree Sourceforge driver. Update the version to match. Signed-off-by:
-
Jacob Keller authored
Instead of using lowercase vlan, vid, or VID, always use VLAN or VLAN ID in comments when referring to VLANs. The original driver code was consistent, but recent patches have not been as consistent with this naming scheme. Signed-off-by:
-
Jacob Keller authored
Signed-off-by:
-
Jacob Keller authored
Improve code style by removing the unnecessary else block of an if statement which immediately returns. Signed-off-by:
-
Jacob Keller authored
Avoid the use of CamelCase for some variable names that previously slipped through review. Reported-by:
-
Jacob Keller authored
Use the ether_addr_copy function instead of copying byte-by-byte in a for-loop by hand. Reported-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 1GbE Intel Wired LAN Driver Updates 2015-12-12 This series contains updates to e1000, e1000e and igb. Joern Engel fixes up the e1000 driver to reduce scheduler latencies by making the eeprom read/write functions scheduler friendly by using a mutex lock instead of a spin lock. Todd adds code for igb to initialize the 88E1543 PHY properly. Then fixed igb to use the correct i210 register for EEMNGCTL, since the i210 has two EEPROM access registers (EEARBC and EEMNGCTL). Dmitry Vyukov provides a fix for e1000 to resolve a data race found with KernelThreadSanitizer (KTSAN), where no memory barriers were being used when buffers get recycled, so the recycled buffers can be corrupted. So use smp_store_release() to update tx_ring->next_to_clean and smp_load_acquire() to read tx_ring->next_to_clean to properly hand off buffers from e1000_clean_tx_irq() to e1000_xmit_frame(). Jarod Wilson fixes igb so that we do not try to unmap a NULL hw_addr. Then cleaned up array_rd32() so that it uses igb_rd32() the same as rd32() and use io_addr() in more places so that we do not have to call E1000_REMOVED(). Janusz Wolak cleans up the e1000 driver by correcting warnings produced by checkpatch.pl for the driver. Jean Sacren provides several patches with general cleanups for e1000 and e1000e, which include code comment fix-ups and cleanup of local variables not needed. Dmitry Fleytman fixes a possible division by zero in the receive interrupt handler for e1000e when working without adaptive interrupt moderation, which is typically disabled on jumbo MTUs. Raanan increases the timeout of the polling bit due to timing changes to the ME firmware on a platform, so increase the timeout to 300ms. Added initial support for i219-LM, which is a LOM that will be available on systems with the Lewisburg Platform Controller HUB (PCH) chipset. Jan Beulich fixes a NULL dereference in igb, due to the adapter->vf _data being NULL while adapter->vfs_allocated_count is non-zero. ==================== Signed-off-by: -
Raanan Avargil authored
i219-LM (3) is a LOM that will be available on systems with the Lewisburg Platform Controller Hub (PCH) chipset from Intel. This patch provides the initial support for the device. Signed-off-by:
Raanan Avargil <raanan.avargil@intel.com> Tested-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
-
Jarod Wilson authored
Clean up array_rd32 so that it uses igb_rd32 the same as rd32, per the suggestion of Alexander Duyck, and use io_addr in more places, so that we don't have the need to call E1000_REMOVED (which simply looks for a null hw_addr) nearly as much. Signed-off-by:
Jarod Wilson <jarod@redhat.com> Acked-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 40GbE Intel Wired LAN Driver Updates 2015-12-12 This series contains updates to i40e and i40evf only. Jesse fixes some trivial static analyzer warnings where BIT() can be used instead of BIT_ULL(). Mitch fixes the virtual channel interface which was using incorrect semantics to remove MAC addresses and would leave incorrect filters active when using VLANs. Also fixes an issue that when VF's are created, the MAC address defaults to all zeros, indicating to the VF driver that it should use a random MAC address. However, the PF driver was incorrectly adding this zero MAC to the filter table, so check for a good address before adding the default filter. Adds a check to make sure that the Tx and Rx rings actually exist before dereferencing them to free resources. Re-classifies several messages which are really for debugging purposes, especially since the driver can fully recover from any of these. Fixed up the VF version strings to match the PF driver. Anjali adds a virtchnl offload to support the expanded version of TCP/UDP PCTYPES for RSS. Shannon fixes i40e to clean the whole MAC filter list when resetting after an intermediate add or delete push to the firmware. v2: added blank line after variable declaration in patch 9 based on feedback from Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> ==================== Signed-off-by: -
Jan Beulich authored
The combined effect of commits 6423fc34 ("igb: do not re-init SR-IOV during probe") and ceee3450 ("igb: make sure SR-IOV init uses the right number of queues") causes VFs no longer getting set up, leading to NULL pointer dereferences due to the adapter's ->vf_data being NULL while ->vfs_allocated_count is non-zero. The first commit not only neglected the side effect of igb_sriov_reinit() that the second commit tried to account for, but also that of setting IGB_FLAG_HAS_MSIX, without which igb_enable_sriov() is effectively a no-op. Calling igb_{,re}set_interrupt_capability() as done here seems to address this, but I'm not sure whether this is better than sinply reverting the other two commits. Signed-off-by:
Jan Beulich <jbeulich@suse.com> Tested-by:
-
Raanan Avargil authored
Due to timing changes to the ME firmware in Skylake, this timer needs to be increased to 300ms. Signed-off-by:
-
Dmitry Fleytman authored
This patch fixes possible division by zero in receive interrupt handler when working without adaptive interrupt moderation. The adaptive interrupt moderation mechanism is typically disabled on jumbo MTUs. Signed-off-by:
Dmitry Fleytman <dmitry@daynix.com> Signed-off-by:
Leonid Bloch <leonid@daynix.com> Tested-by:
-
Janusz Wolak authored
Signed-off-by:
Janusz Wolak <januszvdm@gmail.com> Acked-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by:
-
Jean Sacren authored
By using goto statement, we can achieve sharing the same exit path so that code duplication could be minimized. Signed-off-by:
Jean Sacren <sakiwit@gmail.com> Tested-by:
-
