- 07 Apr, 2017 20 commits
-
-
Tobias Klauser authored
Instead of using a private copy of struct net_device_stats in struct port_info, use stats from struct net_device. Signed-off-by:
Tobias Klauser <tklauser@distanz.ch> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Yuval Mintz says: ==================== qede: support XDP head adjustments Daniel has brought to my attention the fact that qede is the only driver that currently supports XDP but still fails any program where xdp_adjust_head is set on the bpf_prog. This series is meant to remedy this and align qede with the rest of the driver, making it possible to remove said field. Patch #1 contains a minor cache-saving optimization for latter patches. Patches #2 & #3 address existing issues with the qede implementation [#2 should have been a part of this as it addresses something that's affected by the additional headroom; #3 is simply here for the ride]. Patches #4 & #5 add the necessary logic in driver for ingress headroom, the first adding the infrastrucutre needed for supporting the headroon [as currently qede doesn't support such], and the second removing the existing XDP limitation. ==================== Signed-off-by: -
Mintz, Yuval authored
In case an XDP program is attached, reserve XDP_PACKET_HEADROOM bytes at the beginning of the packet for the program to play with. Modify the XDP logic in the driver to fill-in the missing bits and re-calculate offsets and length after the program has finished running to properly reflect the current status of the packet. We can then go and remove the limitation of not supporting XDP programs where xdp_adjust_head is set. Signed-off-by:
Yuval Mintz <Yuval.Mintz@cavium.com> Signed-off-by:
-
Mintz, Yuval authored
Driver currently doesn't support any headroom; The only 'available' space it has in the head of the buffer is due to the placement offset. In order to allow [later] support of XDP adjustment of headroom, modify the the ingress flow to properly handle a scenario where the packets would have such. Signed-off-by:
-
Mintz, Yuval authored
Current implementation of VFs is very tight in regard to queue resources. VFs support for XDP would require quite a bit of additional infrastructure in qede and qed [sharing of queue-zones between queues, more VF cids, mapping of the doorbell bar, etc.]. For now, prevent XDP programs from being attached to VFs. Signed-off-by:
-
Mintz, Yuval authored
Driver is currently using dma_unmap_single() with the address it passed to device for the purpose of forwarding, but the XDP transmission buffer was originally a page allocated for the rx-queue. The mapped address is likely to differ from the original mapped address due to the placement offset. This difference is going to get even bigger once we support headroom. Cache the original mapped address of the page, and use it for unmapping of the buffer when completion arrives for the XDP forwarded packet. Signed-off-by:
-
Mintz, Yuval authored
Currently, each time an ingress packet is passed to networking stack the driver increments a per-queue SW statistic. As we want to have additional fields in the first cache-line of the Rx-queue struct, change flow so this statistic would be updated once per NAPI run. We will later push the statistic to a different cache line. Signed-off-by:
-
David S. Miller authored
Ursula Braun says: ==================== s390 patches for net-next here are some cleanup patches for drivers/s390/net. V2: respin, now patch "s390/qeth: improve endianness handling" is supposed to apply cleanly to net-next ==================== Signed-off-by: -
Hans Wippel authored
Replace ntohs with endianness conversion for the SKB protocol assignment to avoid an endianness warning reported by sparse. No functional change. Signed-off-by:
Hans Wippel <hwippel@linux.vnet.ibm.com> Signed-off-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by:
-
Hans Wippel authored
Use endianness conversions for SKB protocol assignments and usage to avoid endianness warnings reported by sparse. No functional changes. Signed-off-by:
-
Hans Wippel authored
Avoid endianness warnings reported by sparse by (1) using endianness conversions for assigning and using network packet fields, and (2) removing unnecessary endianness conversions from qeth_l3_rebuild_skb. No functional changes. Signed-off-by:
-
Julian Wiedmann authored
1. options.add_hhlen is set but never used, drop it 2. clean up no longer required forward declarations 3. delete all sorts of unused defines Signed-off-by:
Julian Wiedmann <jwi@linux.vnet.ibm.com> Reviewed-by:
Thomas Richter <tmricht@linux.vnet.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
qeth_qdio_output_handler() is the only caller of qeth_handle_send_error() and doesn't care about the return value. Signed-off-by:
-
Julian Wiedmann authored
The ac fields are bitmaps, so format them as hex. While at it, also print the ac2 field. Signed-off-by:
-
Julian Wiedmann authored
better use the constant definitions. Signed-off-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 2017-04-06 This series contains updates to i40e and i40evf. Preethi adds support for the outer checksum and TSO offloads for encapsulated packets for the VF. Mitch fixes a possible memory leak, where we need to remove the client instance when the driver unloads. Also we need to check to see if the client (i40iw) is already present during probe, and add a client instance if necessary. Lastly, make sure we close any attached clients when the driver is removed or shut down to prevent a kernel panic. ==================== Signed-off-by: -
Mitch Williams authored
When the driver is removed or shut down, close any attached clients (i.e. i40iw). This prevents a panic seen sometimes on forced driver removal or system shutdown when iWarp is running. Change-ID: I4f6161e5a73ffbb2fd5883567b007310302bfcb5 Signed-off-by:
Mitch Williams <mitch.a.williams@intel.com> Tested-by:
Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Mitch Williams authored
In some cases, a client (i40iw) may already be present when probe is called. Check for this, and add a client instance if necessary. Change-ID: I2009312694b7ad81f1023919e4c6c86181f21689 Signed-off-by:
-
Mitch Williams authored
When the driver is unloaded, we need to remove the client instance, otherwise we leak memory. Change-ID: If1e7882ac1f6ce15d004722fafbe31afbe0adc9a Signed-off-by:
-
Preethi Banala authored
This patch adds a capability negotiation between VF and PF using ENCAP/ ENCAP_CSUM offload flags in order for the VF to support outer checksum and TSO offloads for encapsulated packets. These capabilities were assumed by default and enabled in current hardware. Going forward, these features needs to be negotiated with PF before advertising to the stack. Additionally, strip out the mac.type checks for X722 since outer checksums are enabled based on the ENCAP_CSUM offload negotiation flag and maintain consistency between drivers in how the features are configured. Change-ID: Ie380a6f57eca557a2bb575b66b12fae36d308920 Signed-off-by:
Preethi Banala <preethi.banala@intel.com> Signed-off-by:
Alan Brady <alan.brady@intel.com> Signed-off-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by:
-
- 06 Apr, 2017 20 commits
-
-
David S. Miller authored
Benjamin Herrenschmidt says: ==================== ftgmac: Rework batch 2 - RX path This is the second batch of updates to the ftgmac100 driver. This one tackles the RX path of the driver, simplifying it greatly to match common practice while significantly increasing the performance. (The bulk of the performance gains of my series will be provided by the TX path improvements, notably fragmented sends, these will be in the next batch). ==================== Signed-off-by: -
Benjamin Herrenschmidt authored
The HW incorrectly calculates the frame size without the vlan tag and compares that against 64. It will thus flag 64-bytes frames with a vlan tag as 60-bytes frames "runt" packets which we'll then drop. Thus we end up dropping ARP packets on vlan's ... It does that whether vlan tag stripping is enabled or not. This works around it by ignoring the "runt" error bit of the frame has been vlan tagged and is at least 60 bytes. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
Benjamin Herrenschmidt authored
Directly access the fields when needed. The accessors add clutter not clarity and in some cases cause unnecessary read-modify-write type access on the slow (uncached) descriptor memory. Signed-off-by:
-
Benjamin Herrenschmidt authored
Signed-off-by:
-
Benjamin Herrenschmidt authored
The current driver receive path allocates pages and stashes them into SKB fragments. This is not particularly useful as we don't support jumbo frames (which wouldn't be great with the small FIFOs on all the known implementations) anyway. It also makes us flush the caches and allocate more memory for RX than necessary. So set our RX buf to our max packet size instead (which we bump to 1536 bytes to account for packets with vlan tags etc...) like most other ethernet drivers. Then allocate skbs when populating the receive ring and DMA directly into them. This simplifies the RX path further. Signed-off-by:
-
Benjamin Herrenschmidt authored
We don't handle fragmented RX packets, so the "looping" helpers to locate the first segment of a packet or to drop a packet aren't actually helping. Take them out and simplify ftgmac100_rx_packet() further as a result. Signed-off-by:
-
Benjamin Herrenschmidt authored
The fast path has a single unlikely() test for any error bit, calling into a helper that sets the appropriate statistics. The various netdev_info aren't particularly interesting. If we want to differentiate the various length errors later we can introduce driver specific stats using ethtool. Signed-off-by:
-
Benjamin Herrenschmidt authored
Read the descriptor field only once and check for IP header checksum errors as well Signed-off-by:
-
Benjamin Herrenschmidt authored
We can occasionally fail to allocate new RX buffers at runtime or when starting the driver. At the moment the latter just fails to open which is fine but the former leaves stale DMA pointers in the ring. Instead, use a scratch page and have all RX ring descriptors point to it by default unless a proper buffer can be allocated. It will help later on when re-initializing the whole ring at runtime on link changes since there is no clean failure path there unlike open(). Signed-off-by:
-
Benjamin Herrenschmidt authored
We don't support jumbo frames, we will never receive a fragmented packet, the RX buffer is always big enough, if not then it's a runaway packet that can be dropped. So take out the loop that handles such things in ftgmac100_rx_packet() which will help with subsequent simplifications and improvements to the RX path Signed-off-by:
-
Benjamin Herrenschmidt authored
Avoids a forward declaration Signed-off-by:
-
git://git.open-mesh.org/linux-mergeDavid S. Miller authored
Simon Wunderlich says: ==================== This feature/cleanup patchset includes the following patches: - bump version strings, by Simon Wunderlich - Code and Style cleanups, by Sven Eckelmann (5 patches) - Remove an unneccessary memset, by Tobias Klauser - DAT and BLA optimizations for various corner cases, by Andreas Pape (5 patches) - forward/rebroadcast packet restructuring, by Linus Luessing (2 patches) - ethtool cleanup and remove unncessary code, by Sven Eckelmann (4 patches) - use net_device_stats from net_device instead of private copy, by Tobias Klauser ==================== Signed-off-by: -
David S. Miller authored
Yuval Mintz says: ==================== qed: Misc cleanups and fixes Patches #1 and #2 revolve around register access performed by driver; The first merely adds some debug, while the second does some fixing of incorrect PTT usage as well as preventing issues similar to those fixed by 6f437d43 ("qed: Don't use attention PTT for configuring BW"). Patch #3 better configures HW for architecture where cacheline isn't 64B. Patches #4-#8 all affect iSCSI related functionaility - adding statistics information [both to driver & management firmware], passing information on number of resources to qedi, and simplifying the Out-of-order implementation in SW. ==================== Signed-off-by:
-
Michal Kalderon authored
No need to maintain the various open archipelagos as a list - The maximal number of them is known, and we can use the CID as key for random-access into the array. Signed-off-by:
Michal Kalderon <Michal.Kalderon@caviumc.om> Signed-off-by:
-
Mintz, Yuval authored
Management firmware can query for some basic iSCSI-related statistics. Provide those just as we do for other protocols. Signed-off-by:
-
Mintz, Yuval authored
Now that management firmware is capable of telling us the number of CQs available for a given PF, qed needs to communicate the number to qedi so it would know have many to use. Signed-off-by:
-
Mintz, Yuval authored
Firmware provides a statistic for the number of out-of-order isles it used - fill it in the iscsi-related statistics. Signed-off-by:
-
Mintz, Yuval authored
Before initializing the chip's engine, driver currently closes a set of registers on the HW's ingress flow to prevent packets from slipping in while they're not supposed to. This configuration is insufficient, as there are some scenarios where packets would still arrive even when said registers are set, but the management firmware already closes other per-port registers that do suffice, making this setting unnecessray. Signed-off-by:
-
Tomer Tayar authored
Default HW configuration is optimal for an architecture where cache line size is 64B. During chip initialization, properly initialize the cache line size in HW to avoid possible redundant PCI transactions. Signed-off-by:
Tomer Tayar <Tomer.Tayar@cavium.com> Signed-off-by:
-
Rahul Verma authored
In order to access HW registers driver needs to acquire a PTT entry [mapping between bar memory and internal chip address]. Since acquiring PTT entries could fail [at least in theory] as their number is finite and other flows can hold them, we reserve special PTT entries for 'important' enough flows - ones we want to guarantee that would not be susceptible to such issues. One such special entry is the 'main' PTT which is meant to be used in flows such as chip initialization and de-initialization. However, there are other flows that are also using that same entry for their own purpose, and might run concurrently with the original flows [notice that for most cases using the main-ptt by mistake, such a race is still impossible, at least today]. This patch re-organizes the various functions that currently use the main_ptt in one of two ways: - If a function shouldn't use the main_ptt it starts acquiring and releasing it's own PTT entry and use it instead. Notice if those functions previously couldn't fail, they now can [as acquisition might fail]. - Change the prototypes so that the main_ptt would be received as a parameter [instead of explicitly accessing it]. This prevents the future risk of adding codes that introduces new use-cases for flows using the main_ptt, ones that might be in race with the actual 'main' flows. Signed-off-by:Rahul Verma <Rahul.Verma@cavium.com> Signed-off-by:
-
