- 30 Nov, 2009 5 commits
-
-
Ben Hutchings authored
Refactor PHY, MAC and NIC configuration operations so that the existing link configuration can be re-pushed with: efx->phy_op->reconfigure(efx); efx->mac_op->reconfigure(efx); and a new configuration with: efx->nic_op->reconfigure_port(efx); (plus locking and error-checking). We have not held the link settings in software (aside from flow control), and have relied on asking the hardware what they are. This is a problem because in some cases the hardware may no longer be in a state to tell us. In particular, if an entire multi-port board is reset through one port, the driver bindings to other ports have no chance to save settings before recovering. We only actually need to keep track of the autonegotiation settings, so add an ethtool advertising mask to struct efx_nic, initialise it in PHY init and update it as necessary. Remove now-unneeded uses of efx_phy_op::{get,set}_settings() and struct ethtool_cmd. Much of this was done by Steve Hodgson <shodgson@solarflare.com>. Signed-off-by:Ben Hutchings <bhutchings@solarflare.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Ben Hutchings authored
This is preparation for adding differing implementations for new NICs. Signed-off-by:
-
Ben Hutchings authored
Pause frame generation is gated by both RX_XOFF_MAC_EN and an enable bit in each MAC. RX_XOFF_MAC_EN bit always reads back as 0 so we need to set it correctly every time we modify RX_CFG_REG. Simplify this by always setting it to 1 and only changing the enable bits in the MACs. Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
This register only needs to be written after reset, not each time we enable interrupts. Signed-off-by:
-
- 29 Nov, 2009 25 commits
-
-
Eric Dumazet authored
pktgen threads are bound to given CPU, we can allocate memory for these threads in a NUMA aware way. After a pktgen session on two threads, we can check flows memory was allocated on right node, instead of a not related one. # grep pktgen_thread_write /proc/vmallocinfo 0xffffc90007204000-0xffffc90007385000 1576960 pktgen_thread_write+0x3a4/0x6b0 [pktgen] pages=384 vmalloc N0=384 0xffffc90007386000-0xffffc90007507000 1576960 pktgen_thread_write+0x3a4/0x6b0 [pktgen] pages=384 vmalloc N1=384 Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
David S. Miller authored
Conflicts: drivers/ieee802154/fakehard.c drivers/net/e1000e/ich8lan.c drivers/net/e1000e/phy.c drivers/net/netxen/netxen_nic_init.c drivers/net/wireless/ath/ath9k/main.c
-
PJ Waskiewicz authored
This patch extends the ethtool interface to display what PHY is currently connected to a NIC. The results can be viewed in ethtool ethX output. Signed-off-by:
Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
PJ Waskiewicz authored
This patch allows a base driver to specify Direct Attach as the type of port through the ethtool interface. Signed-off-by:
-
Shannon Nelson authored
This fixes an issue when clearing out the RAR entries. If RAR[0] is the only address in use, don't clear the others. Signed-off-by:
Shannon Nelson <shannon.nelson@intel.com> Signed-off-by:
-
Don Skidmore authored
82598 shouldn't try and access LINKS2 while configuring link and flow control. This is an 82599-only register. Signed-off-by:
Don Skidmore <donald.c.skidmore@intel.com> Signed-off-by:
-
PJ Waskiewicz authored
Flow Control autoneg should be disabled for certain adapters that don't support autonegotiation of Flow Control at 10 gigabit. These interfaces are the 10GBASE-T devices, CX4, and SFP+, all running at 10 gigabit only. 1 gigabit is fine. Signed-off-by:
-
Shannon Nelson authored
The driver was doing a divide by zero when adjusting tx-usecs. This patch removes the divide by zero code and changes the logic slightly to ignore tx-usecs in the case of shared TxRx vectors. Cc: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by:
-
andrew hendry authored
Calls to x25_dev_get check for dev = NULL which was not set. It allowed x25 to set routes and ioctls on down interfaces. This caused oopses and refcnt problems on device_unregister. Signed-off-by:
Andrew Hendry <andrew.hendry@gmail.com> Signed-off-by:
-
andrew hendry authored
Adds error checking to x25_init. Signed-off-by:
-
andrew hendry authored
Moves the CONFIG_SYSCTL ifdefs in x25_init into header. Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Kevin Hilman authored
Convert smc91x driver from legacy PM hooks over to using dev_pm_ops. Tested on OMAP3 platform. Signed-off-by:
Kevin Hilman <khilman@deeprootsystems.com> Acked-by:
Nicolas Pitre <nico@fluxnic.net> Signed-off-by:
-
-
Andrei Pelinescu-Onciul authored
When retransmitting due to T3 timeout, retransmit all the in-flight chunks for the corresponding transport/path, including chunks sent less then 1 rto ago. This is the correct behaviour according to rfc4960 section 6.3.3 E3 and "Note: Any DATA chunks that were sent to the address for which the T3-rtx timer expired but did not fit in one MTU (rule E3 above) should be marked for retransmission and sent as soon as cwnd allows (normally, when a SACK arrives). ". This fixes problems when more then one path is present and the T3 retransmission of the first chunk that timeouts stops the T3 timer for the initial active path, leaving all the other in-flight chunks waiting forever or until a new chunk is transmitted on the same path and timeouts (and this will happen only if the cwnd allows sending new chunks, but since cwnd was dropped to MTU by the timeout => it will wait until the first heartbeat). Example: 10 packets in flight, sent at 0.1 s intervals on the primary path. The primary path is down and the first packet timeouts. The first packet is retransmitted on another path, the T3 timer for the primary path is stopped and cwnd is set to MTU. All the other 9 in-flight packets will not be retransmitted (unless more new packets are sent on the primary path which depend on cwnd allowing it, and even in this case the 9 packets will be retransmitted only after a new packet timeouts which even in the best case would be more then RTO). This commit reverts d0ce9291 and also removes the now unused transport->last_rto, introduced in b6157d8e. p.s The problem is not only when multiple paths are there. It can happen in a single homed environment. If the application stops sending data, it possible to have a hung association. Signed-off-by:
Andrei Pelinescu-Onciul <andrei@iptel.org> Signed-off-by:
Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by:
-
Ben Hutchings authored
Add 'likely' hint to test of rx_checksum_enabled. Don't count IP fragments; the IP stack can do that. Do count non-matching multicast packets. Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Matthew Slattery authored
The "bug9141 workaround" of setting TX_FLUSH_MIN_LEN_EN should really be considered as a normal bit of configuration rather than a workaround. Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Steve Hodgson authored
The strap bits are only important on Falcon A and all production boards using it have fixed-speed 10G PHYs. Replace dummy MAC operations with default MAC operations. Signed-off-by:
-
Steve Hodgson authored
We never use MDIO in atomic context, so we don't need to spin. Signed-off-by:
-
Steve Hodgson authored
Reading standard registers on the QT2025C before its firmware has booted may cause the boot process to fail. Therefore, follow the recommended reset sequence before reading its id registers. Either order works for the QT2022C2, so don't differentiate. Signed-off-by:
-
Steve Hodgson authored
Falcon can generate events for LASI interrupts from the PHY, but in practice we have never implemented this in reference designs. Instead we have polled, inserted the appropriate events, and then handled the events later. This is a waste of time and code. Instead, make PHY poll functions update the link state synchronously and report whether it changed. We can still make use of the LASI registers as a shortcut on the SFT9001. Signed-off-by:
-
- 27 Nov, 2009 1 commit
-
-
Patrick McHardy authored
Currently the UP/DOWN state of VLANs is synchronized to the state of the underlying device, meaning all VLANs are set down once the underlying device is set down. This causes all routes to the VLAN devices to vanish. Add a flag to specify a "loose binding" mode, in which only the operstate is transfered, but the VLAN device state is independant. Signed-off-by:
Patrick McHardy <kaber@trash.net> Signed-off-by:
-
- 26 Nov, 2009 9 commits
-
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
From: Steve Hodgson <shodgson@solarflare.com> Currently we can set multicast hash immediately (in atomic context) but must delay setting MAC promiscuity. There is not that much point in deferring one but not the other, and setting the multicast hash on Siena will involve a firmware request. So process them both in efx_mac_work(). Also, set the broadcast bit in the multicast hash in efx_set_multicast_list(), since this is required for both Falcon and Siena. Signed-off-by:
-
Ben Hutchings authored
From: Steve Hodgson <shodgson@solarflare.com> Only the XMAC on Falcon needs help from the driver to poll and reset the MAC-PHY link (XAUI); GMII is a simple parallel bus and on later NICs firmware takes care of the XAUI link. Also, an XMAC interrupt currently schedules a work item which simply clears a flag (efx_nic::mac_up) to be checked by the regular monitor (or the next link reconfiguration, if that is sooner). Rename the flag to xmac_poll_required, changing its sense. Remove the needless indirection and just set the flag immediately. Call falcon_xmac_poll() directly where required. Add a new generic operation mac_op::check_fault to check the link outside of regular monitoring, as required during self-tests. (Note that this leaves us with an unused work item, but we will immediately have another use for it.) Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
Ben Hutchings authored
From: Steve Hodgson <shodgson@solarflare.com> Currently we initiate MAC stats DMA and busy-wait for completion when stats are requested. We can improve on this with a periodic timer to initiate and poll for stats, and opportunistically poll when stats are requested. Since efx_nic::stats_disable_count and efx_stats_{disable,enable}() are Falcon-specific, rename them and move them accordingly. Signed-off-by: -
Ben Hutchings authored
Although efx_init_port() is only called at probe time and so cannot race with port reconfiguration, most of the functions it calls can expect to be called with the MAC lock held. Signed-off-by:
-
Steve Hodgson authored
Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
-
