Add logic to calculate how many shared or non shared rings can be
supported. Default is to use shared rings.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to support dedicated or shared completion rings, the ring
indexing and mapping are re-structured as below:
1. bp->grp_info[] array index is 1:1 with bp->bnapi[] array index and
completion ring index.
2. rx rings 0 to n will be mapped to completion rings 0 to n.
3. If tx and rx rings share completion rings, then tx rings 0 to m will
be mapped to completion rings 0 to m.
4. If tx and rx rings use dedicated completion rings, then tx rings 0 to
m will be mapped to completion rings n + 1 to n + m.
5. Each tx or rx ring will use the corresponding completion ring index
for doorbell mapping and MSIX mapping.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Each bnxt_napi structure may no longer be having both an rx ring and
a tx ring. Check for a valid ring before using it.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, an rx and a tx ring are always paired with a completion ring.
We want to restructure it so that it is possible to have a dedicated
completion ring for tx or rx only.
The bnxt hardware uses a completion ring for rx and tx events. The driver
has to process the completion ring entries sequentially for the rx and tx
events. Using a dedicated completion ring for rx only or tx only has these
benefits:
1. A burst of rx packets can cause delay in processing tx events if the
completion ring is shared. If tx queue is stopped by BQL, this can cause
delay in re-starting the tx queue.
2. A completion ring is sized according to the rx and tx ring size rounded
up to the nearest power of 2. When the completion ring is shared, it is
sized by adding the rx and tx ring sizes and then rounded to the next power
of 2, often with a lot of wasted space.
3. Using dedicated completion ring, we can adjust the tx and rx coalescing
parameters independently for rx and tx.
The first step is to separate the rx and tx ring structures from the
bnxt_napi struct.
In this patch, an rx ring and a tx ring will point to the same bnxt_napi
struct to share the same completion ring. No change in ring assignment
and mapping yet.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By adding 3 separate functions to dump the different ring states.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added the PCI IDs for the BCM57301 and BCM57402 controllers.
Signed-off-by: David Christensen <davidch@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This interface will be forward compatible with future changes.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer firmware will return the ring group resource when we call
hwrm_func_qcaps(). To be compatible with older firmware, use the
number of tx rings as the number of ring groups if the older firmware
returns 0. When determining how many rx rings we can support, take
the ring group resource in account as well in _bnxt_get_max_rings().
Divide and assign the ring groups to VFs.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We need to keep track of all resources, such as rx rings, tx rings,
cmpl rings, rss contexts, stats contexts, vnics, after we have
divided them for the VFs. Otherwise, subsequent ring changes on
the PF may not work correctly.
We adjust all max resources in struct bnxt_pf_info after they have been
assigned to the VFs. There is no need to keep the separate
max_pf_tx_rings and max_pf_rx_rings.
When SR-IOV is disabled, we call bnxt_hwrm_func_qcaps() to restore the
max resources for the PF.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
1. Use local variable pf for repeated access to this pointer.
2. The 2nd argument num_vfs was unnecessarily declared as pointer to int.
This function doesn't change num_vfs so change the argument to int.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The hardware resources required to enable NTUPLE varies depending on
how many rx channels are configured. We need to make sure we have the
resources before we enable NTUPLE. Add bnxt_rfs_capable() to do the
checking.
In addition, we need to do the same checking in ndo_fix_features(). As
the rx channels are changed using ethtool -L, we call
netdev_update_features() to make the necessary adjustment for NTUPLE.
Calling netdev_update_features() in netif_running() state but before
calling bnxt_open_nic() would be a problem. To make this work,
bnxt_set_features() has to be modified to test for BNXT_STATE_OPEN for
the true hardware state instead of checking netif_running().
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If hardware completes single segment rx frames, don't bother setting
up all the GRO related fields. Pass the SKB up as a normal frame.
Reviewed-by: vasundhara volam <vvolam@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Also, no need to check for bp->rx_nr_rings as it is always >= 1. If the
allocation fails, it is not a fatal error and we can still proceed.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rx_l4_csum_error is now incremented only when offload is enabled
Signed-off-by: Satish Baddipadige <sbaddipa@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NC-SI firmware of type apeFW (10) is now supported.
Signed-off-by: Rob Swindell <swindell@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove the unnecessary "if" statement before the "for" statement:
if (x) {
for (i = 0; i < x; i++)
...
}
Also, change the ring free function to return void as it only returns 0.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
During remove_one, the driver should issue hwrm_func_drv_unrgtr
command to inform firmware that this function has been unloaded.
This is to let firmware keep track of driver present/absent state
when driver is gracefully unloaded. A keep alive timer is needed
later to keep track of driver state during abnormal shutdown.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
FW has a rare corner case in which a fragmented packet using lots
of frags would not be linearized, causing the FW to assert while trying
to transmit the packet.
To prevent this, we need to make sure the window of fragements containing
MSS worth of data contains 1 BD less than for regular packets due to
the additional parsing BD.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/geneve.c
Here we had an overlapping change, where in 'net' the extraneous stats
bump was being removed whilst in 'net-next' the final argument to
udp_tunnel6_xmit_skb() was being changed.
Signed-off-by: David S. Miller <davem@davemloft.net>
The reset logic calls bnxt_close_nic() and bnxt_open_nic() under rtnl_lock
from bnxt_sp_task. BNXT_STATE_IN_SP_TASK must be cleared before calling
bnxt_close_nic() to avoid deadlock.
v2: Fixed white space error. Thanks Dave.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When implementing driver reset from tx_timeout in the next patch,
bnxt_close_nic() will be called from the sp_task workqueue. Calling
cancel_work() on sp_task will hang the workqueue.
Instead, set a new bit BNXT_STATE_IN_SP_TASK when bnxt_sp_task() is running.
bnxt_close_nic() will wait for BNXT_STATE_IN_SP_TASK to clear before
proceeding.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This allows multiple independent bits to be set for various states.
Subsequent patches to implement tx timeout reset will require this.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The declaration of the bitmap vf_req_snif_bmap using fixed array of
unsigned long will only work on 64-bit archs. Use DECLARE_BITMAP instead
which will work on all archs.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These fields are updated but never read.
Remove the overhead.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Under heavy TX load, bnx2x_poll() can loop forever and trigger
soft lockup bugs.
A napi poll handler must yield after one TX completion round,
risk of livelock is too high otherwise.
Bug is very easy to trigger using a debug build, and udp flood, because
of added cpu cycles in TX completion, and we do not receive enough
packets to break the loop.
Reported-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ariel Elior <ariel.elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The 'flags' field in bnx2x_stats_arr[] serves only one purpose - to tell
us if the statistic is a per-port stat and thus should not be shown for
virtual functions. It's strange that the field can have three different
values. A boolean will do just fine.
Also remove IS_FUNC_STAT(). It was used only once and it's in fact just
a negation of IS_PORT_STAT().
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It's supposed to be impossible for TPA to give us anything else
than IPv4 or IPv6 here. But in case there is a way to reach this error
by some strange received frames, we don't want to flood the kernel log.
WARN_ONCE is better for this purpose.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
alloc_pages() already prints a warning when it fails. No need to emit
another message. Certainly not at KERN_ERR level, because it is no big
deal if this GFP_ATOMIC allocation fails occasionally.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/renesas/ravb_main.c
kernel/bpf/syscall.c
net/ipv4/ipmr.c
All three conflicts were cases of overlapping changes.
Signed-off-by: David S. Miller <davem@davemloft.net>
These new helpers simplify implementing multi-driver modules and
properly handle failure to register one driver by unregistering all
previously registered drivers.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Call bnxt_cfg_rx_mode() in bnxt_init_chip() to setup uc_list and
mc_list mac address filters. Before the patch, uc_list is not
setup again after chip reset (such as ethtool ring size change)
and macvlans don't work any more after that.
Modify bnxt_cfg_rx_mode() to return error codes appropriately so
that the init chip sequence can detect any failures.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For PF, the bp->pf.mac_addr always holds the permanent MAC
addr assigned by the HW. For VF, the bp->vf.mac_addr always
holds the administrator assigned VF MAC addr. The random
generated VF MAC addr should never get stored to bp->vf.mac_addr.
This way, when the VF wants to change the MAC address, we can tell
if the adminstrator has already set it and disallow the VF from
changing it.
v2: Fix compile error if CONFIG_BNXT_SRIOV is not set.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The existing ndo_set_mac_address only copies the new MAC addr
and didn't set the new MAC addr to the HW. The correct way is
to delete the existing default MAC filter from HW and add
the new one. Because of RFS filters are also dependent on the
default mac filter l2 context, the driver must go thru
close_nic() to delete the default MAC and RFS filters, then
open_nic() to set the default MAC address to HW.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 46e8a249423ff "bnx2x: Add FW 7.13.1.0" added said .bin FW to
linux-firmware; This patch incorporates the FW in the bnx2x driver.
This introduces 2 fixes/enhancements:
- In some management protocols there are outer-vlan configurations
that can be dynamically changed while device is running. This fixes
some corner cases where such a change did not take effect.
- Prevent VFs from sending MAC control frames; FW would treat a VF
sending such a packet as malicious and block any further communication
done by the VF.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Today, port statistics are being presented when using `ethool -S' only
for single-function devices, but there are some port statistics which are
crucial for analyzing bottle-necks. E.g., HW Rx discards due to lack of
buffer space [when device isn't handling ingress traffic fast enough].
Judging the pros and cons, it was decided that in-order to better support
automatic dump-gathering tools, bnx2x should no longer hide those stats.
This leaves only VFs lacking the port statistics.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Driver already has an internal counter for number of times a given queue
had to be stopped due to Tx ring exhaustion.
This add the counter to the statistics presented by driver, e.g., by using
`ethtool -S'.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commmit ac7eccd4d4 "bnx2x: track vxlan port count" contains a bug -
Instead of achieving the required goal, vxlan configuration would not
be removed since we're decrementing the port instead of the counter.
CC: Jiri Benc <jbenc@redhat.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NAPI drivers no longer need to observe a particular protocol
to benefit from busy polling (CONFIG_NET_RX_BUSY_POLL=y)
napi_hash_add() and napi_hash_del() are automatically called
from core networking stack, respectively from
netif_napi_add() and netif_napi_del()
This patch depends on free_netdev() and netif_napi_del() being
called from process context, which seems to be the norm.
Drivers might still prefer to call napi_hash_del() on their
own, since they might combine all the rcu grace periods into
a single one, knowing their NAPI structures lifetime, while
core networking stack has no idea of a possible combining.
Once this patch proves to not bring serious regressions,
we will cleanup drivers to either remove napi_hash_del()
or provide appropriate rcu grace periods combining.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
netif_tx_napi_add() is a variant of netif_napi_add()
It should be used by drivers that use a napi structure
to exclusively poll TX.
We do not want to add this kind of napi in napi_hash[] in following
patches, adding generic busy polling to all NAPI drivers.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We would like to automatically provide busy polling support
to all NAPI drivers, without them having to implement anything.
skb_mark_napi_id() can be called from napi_gro_receive() and
napi_get_frags().
Few drivers are still calling skb_mark_napi_id() because
they use netif_receive_skb(). They should eventually call
napi_gro_receive() instead. I will leave this to drivers
maintainers.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 05cc5a39dd "bnx2x: add vlan filtering offload" introduced
a regression in regard for vlans for 57710, 57711 adapters -
Loading 8021q module on a machine with such an adapter would cause
a null pointer dereference, as the driver mistakenly publishes it
has capabilities for vlan CTAG filtering.
Reported-by: Otto Sabart <osabart@redhat.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
1) Fix null deref in xt_TEE netfilter module, from Eric Dumazet.
2) Several spots need to get to the original listner for SYN-ACK
packets, most spots got this ok but some were not. Whilst covering
the remaining cases, create a helper to do this. From Eric Dumazet.
3) Missiing check of return value from alloc_netdev() in CAIF SPI code,
from Rasmus Villemoes.
4) Don't sleep while != TASK_RUNNING in macvtap, from Vlad Yasevich.
5) Use after free in mvneta driver, from Justin Maggard.
6) Fix race on dst->flags access in dst_release(), from Eric Dumazet.
7) Add missing ZLIB_INFLATE dependency for new qed driver. From Arnd
Bergmann.
8) Fix multicast getsockopt deadlock, from WANG Cong.
9) Fix deadlock in btusb, from Kuba Pawlak.
10) Some ipv6_add_dev() failure paths were not cleaning up the SNMP6
counter state. From Sabrina Dubroca.
11) Fix packet_bind() race, which can cause lost notifications, from
Francesco Ruggeri.
12) Fix MAC restoration in qlcnic driver during bonding mode changes,
from Jarod Wilson.
13) Revert bridging forward delay change which broke libvirt and other
userspace things, from Vlad Yasevich.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (65 commits)
Revert "bridge: Allow forward delay to be cfgd when STP enabled"
bpf_trace: Make dependent on PERF_EVENTS
qed: select ZLIB_INFLATE
net: fix a race in dst_release()
net: mvneta: Fix memory use after free.
net: Documentation: Fix default value tcp_limit_output_bytes
macvtap: Resolve possible __might_sleep warning in macvtap_do_read()
mvneta: add FIXED_PHY dependency
net: caif: check return value of alloc_netdev
net: hisilicon: NET_VENDOR_HISILICON should depend on HAS_DMA
drivers: net: xgene: fix RGMII 10/100Mb mode
netfilter: nft_meta: use skb_to_full_sk() helper
net_sched: em_meta: use skb_to_full_sk() helper
sched: cls_flow: use skb_to_full_sk() helper
netfilter: xt_owner: use skb_to_full_sk() helper
smack: use skb_to_full_sk() helper
net: add skb_to_full_sk() helper and use it in selinux_netlbl_skbuff_setsid()
bpf: doc: correct arch list for supported eBPF JIT
dwc_eth_qos: Delete an unnecessary check before the function call "of_node_put"
bonding: fix panic on non-ARPHRD_ETHER enslave failure
...
__GFP_WAIT has been used to identify atomic context in callers that hold
spinlocks or are in interrupts. They are expected to be high priority and
have access one of two watermarks lower than "min" which can be referred
to as the "atomic reserve". __GFP_HIGH users get access to the first
lower watermark and can be called the "high priority reserve".
Over time, callers had a requirement to not block when fallback options
were available. Some have abused __GFP_WAIT leading to a situation where
an optimisitic allocation with a fallback option can access atomic
reserves.
This patch uses __GFP_ATOMIC to identify callers that are truely atomic,
cannot sleep and have no alternative. High priority users continue to use
__GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and
are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify
callers that want to wake kswapd for background reclaim. __GFP_WAIT is
redefined as a caller that is willing to enter direct reclaim and wake
kswapd for background reclaim.
This patch then converts a number of sites
o __GFP_ATOMIC is used by callers that are high priority and have memory
pools for those requests. GFP_ATOMIC uses this flag.
o Callers that have a limited mempool to guarantee forward progress clear
__GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall
into this category where kswapd will still be woken but atomic reserves
are not used as there is a one-entry mempool to guarantee progress.
o Callers that are checking if they are non-blocking should use the
helper gfpflags_allow_blocking() where possible. This is because
checking for __GFP_WAIT as was done historically now can trigger false
positives. Some exceptions like dm-crypt.c exist where the code intent
is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to
flag manipulations.
o Callers that built their own GFP flags instead of starting with GFP_KERNEL
and friends now also need to specify __GFP_KSWAPD_RECLAIM.
The first key hazard to watch out for is callers that removed __GFP_WAIT
and was depending on access to atomic reserves for inconspicuous reasons.
In some cases it may be appropriate for them to use __GFP_HIGH.
The second key hazard is callers that assembled their own combination of
GFP flags instead of starting with something like GFP_KERNEL. They may
now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless
if it's missed in most cases as other activity will wake kswapd.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Vitaly Wool <vitalywool@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Instead of always calling pci_sriov_disable() in remove_one(),
the driver should detect whether VFs are currently assigned
to the VMs. If the VFs are active in VMs, then it should not
disable SRIOV as it is catastrophic to the VMs. Instead,
it just leaves the VFs alone and continues to unload the PF.
The user can then cleanup the VMs even after the PF driver
has been unloaded.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Assign the return value from bitmap_find_free_region() to an integer
variable and check for negative error codes first, before assigning
the bit ID to the unsigned sw_id field.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to use offset 0x4014 for reading CAG interrupt status,
the actual CAG register must be mapped to GRC bar0 window #4.
Otherwise, the driver is reading garbage. This patch corrects
this issue.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The profile ID in the completion record needs to be ANDed with the
profile ID mask of 0x1f. This bug was causing the SKB hash type
and the gso_type to be wrong in some cases.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the sp event bits to be bit positions instead of bit values since
the bit helper functions are expecting the former.
Signed-off-by: Jeffrey Huang <huangjw@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
