From: Sunil Goutham
This patch changes the way VF's irqs are visible in /proc/interrupts.
Instead of VF id, logical interface's netdev name is used for IRQ
naming and also all secondary VF's interrupts in multiqset config
use primary VF's netdev name.
Signed-off-by: Sunil Goutham
---
drivers/n
From: Jerin Jacob
When SQ/TXQ is reclaimed i.e reset it's stats also automatically reset
by HW. This is not the case with RQ. Also VF doesn't have write access
to statistics counter registers. Hence a new Mbox msg is introduced which
supports resetting RQ, SQ and full Qset stats. Currently only R
From: Sunil Goutham
Instead of a round about way of converting buffers to SKBs and
combining them into a frag list, use standard skb_add_rx_frag()
API to merge page fragments. This code is useful when incoming
packets are of size more than RCV_FRAG_LEN which is currently
set to 2048bytes.
Signed
From: Radoslaw Biernacki
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++--
drivers/net/ethernet/cavium/thunder/nicvf_m
From: Sunil Goutham
Set MAC addresses only for primary VF's and don't for
secondary VFs.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 2 +-
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --g
From: Sunil Goutham
This patch adds RGX/RGMII interface type support to BGX
driver. This type of interface is supported by 81xx SOC.
CN81XX VNIC has 8 VFs and max possible LMAC interfaces are 9,
hence RGMII interface will not work if all DLMs are in BGX mode
and all 8 LMACs are enabled
Signed-o
From: Sunil Goutham
81xx has only 4 CPUs, so it doesn't make sense to initialize
entire Qset i.e 8 queues by default. Made changes to queue
initialization to init queues equal to number of CPUs or
8 queues whichever is lesser. Also this will be applicable to
VMs with VNIC VF attached and having l
From: Sunil Goutham
Unlike 88xx, CQE_RX descriptor's tunnelling extension i.e CQE_RX2_S
is always enabled on 81xx/83xx and HW does insert these fields into
CQE_RX. As a result receive buffer addresses will now be present at
7th word of CQE_RX instead of 6th.
Enable CQE_RX2_S on 88xx pass 2.x as
From: Sunil Goutham
On 88xx all LMACs in a BGX will be in same mode but on 81xx
BGX can be split as two and there can be LMACs configured in
different modes.
These changes move lmac_type, lane2serdes fields into per lmac
struct from BGX struct. Got rid of qlm_mode field which has become
redundan
From: Sunil Goutham
Current driver has most of the HW maximums info like no of channels,
traffic limiters, RSS indices e.t.c in the form of macros. These have
been moved into a 'hw_info' structure so that support for VNIC on
newer chips with different set of HW maximums can be added.
Signed-off-
From: Sunil Goutham
This patch series adds support for VNIC on 81xx and 83xx SOCs.
81xx/83xx is different from 88xx in terms of capabilities and different
types of interfaces supported (eg: QSGMII, RGMII) and have DLMs instead
of QLMs which allows single BGX to have interfaces of different LMAC t
From: Sunil Goutham
This patch adds PCI device IDs of VNIC on newer chips and also
registers VF driver with them. Device id remains same for all
versions of chips but subsystem device id changes.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h| 10 +-
Hi David,
I have found that I missed some piece of code in a patch
and also discovered a corner case issue while further testing.
Please ignore this patch set, will submit another version soon.
Thanks,
Sunil.
On Wed, Jul 13, 2016 at 11:29 AM, wrote:
> From: Sunil Goutham
>
> This patch serie
From: Sunil Goutham
Instead of a round about way of converting buffers to SKBs and
combining them into a frag list, use standard skb_add_rx_frag()
API to merge page fragments. This code is useful when incoming
packets are of size more than RCV_FRAG_LEN which is currently
set to 2048bytes.
Signed
From: Sunil Goutham
Current driver has most of the HW maximums info like no of channels,
traffic limiters, RSS indices e.t.c in the form of macros. These have
been moved into a 'hw_info' structure so that support for VNIC on
newer chips with different set of HW maximums can be added.
Signed-off-
From: Sunil Goutham
On 88xx all LMACs in a BGX will be in same mode but on 81xx
BGX can be split as two and there can be LMACs configured in
different modes.
These changes move lmac_type, lane2serdes fields into per lmac
struct from BGX struct. Got rid of qlm_mode field which has become
redundan
From: Sunil Goutham
This patch adds PCI device IDs of VNIC on newer chips and also
registers VF driver with them. Device id remains same for all
versions of chips but subsystem device id changes.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h| 10 +-
From: Radoslaw Biernacki
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++--
drivers/net/ethernet/cavium/thunder/nicvf_m
From: Sunil Goutham
This patch adds support for QSGMII interface type to
the BGX driver. This type of interface is supported by
81xx SOC.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 55 ++-
drivers/net/ethernet/cavium/thunder/thunder
From: Jerin Jacob
When SQ/TXQ is reclaimed i.e reset it's stats also automatically reset
by HW. This is not the case with RQ. Also VF doesn't have write access
to statistics counter registers. Hence a new Mbox msg is introduced which
supports resetting RQ, SQ and full Qset stats. Currently only R
From: Sunil Goutham
This patch adds info on HW maximums of 81xx/83xx and also
configures receive and transmit datapaths accordingly.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c| 87 ++-
drivers/net/ethernet/cavium/thunder/nic_reg.h
From: Sunil Goutham
This patch adds RGX/RGMII interface type support to BGX
driver. This type of interface is supported by 81xx SOC.
CN81XX VNIC has 8 VFs and max possible LMAC interfaces are 9,
hence RGMII interface will not work if all DLMs are in BGX mode
and all 8 LMACs are enabled.
Signed-
From: Sunil Goutham
Set MAC addresses only for primary VF's and don't for
secondary VFs.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 2 +-
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --g
From: Sunil Goutham
This patch series adds support for VNIC on 81xx and 83xx SOCs.
81xx/83xx is different from 88xx in terms of capabilities and different
types of interfaces supported (eg: QSGMII, RGMII) and have DLMs instead
of QLMs which allows single BGX to have interfaces of different LMAC t
From: Sunil Goutham
This patch changes the way VF's irqs are visible in /proc/interrupts.
Instead of VF id, logical interface's netdev name is used for IRQ
naming and also all secondary VF's interrupts in multiqset config
use primary VF's netdev name.
Signed-off-by: Sunil Goutham
---
drivers/n
From: Sunil Goutham
This patch adds support for BGX module on 81xx where a BGX
can be split and have different LMACs configured in
different modes.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 112 --
drivers/net/ethernet/cavium/thund
From: Sunil Goutham
Unlike 88xx, CQE_RX descriptor's tunnelling extension i.e CQE_RX2_S
is always enabled on 81xx/83xx and HW does insert these fields into
CQE_RX. As a result receive buffer addresses will now be present at
7th word of CQE_RX instead of 6th.
Enable CQE_RX2_S on 88xx pass 2.x as
From: Sunil Goutham
88xx has 128 VFs, 81xx has 8 VFs and 83xx will have 32VFs.
Made changes to PF driver such that mailbox interrupt enable
registers are configuired based on number of VFs HW supports.
Also cleanedup mailbox irq handler registration code.
Signed-off-by: Sunil Goutham
---
drive
From: Sunil Goutham
81xx has only 4 CPUs, so it doesn't make sense to initialize
entire Qset i.e 8 queues by default. Made changes to queue
initialization to init queues equal to number of CPUs or
8 queues whichever is lesser. Also this will be applicable to
VMs with VNIC VF attached and having l
From: Sunil Goutham
88xx has 128 VFs, 81xx has 8 VFs and 83xx will have 32VFs.
Made changes to PF driver such that mailbox interrupt enable
registers are configuired based on number of VFs HW supports.
Also cleanedup mailbox irq handler registration code.
Signed-off-by: Sunil Goutham
---
drive
From: Sunil Goutham
This patch changes the way VF's irqs are visible in /proc/interrupts.
Instead of VF id, logical interface's netdev name is used for IRQ
naming and also all secondary VF's interrupts in multiqset config
use primary VF's netdev name.
Signed-off-by: Sunil Goutham
---
drivers/n
From: Sunil Goutham
This patch adds support for QSGMII interface type to
the BGX driver. This type of interface is supported by
81xx SOC.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 55 ++-
drivers/net/ethernet/cavium/thunder/thunder
From: Sunil Goutham
Current driver has most of the HW maximums info like no of channels,
traffic limiters, RSS indices e.t.c in the form of macros. These have
been moved into a 'hw_info' structure so that support for VNIC on
newer chips with different set of HW maximums can be added.
Signed-off-
From: Sunil Goutham
81xx has only 4 CPUs, so it doesn't make sense to initialize
entire Qset i.e 8 queues by default. Made changes to queue
initialization to init queues equal to number of CPUs or
8 queues whichever is lesser. Also this will be applicable to
VMs with VNIC VF attached and having l
From: Sunil Goutham
Unlike 88xx, CQE_RX descriptor's tunnelling extension i.e CQE_RX2_S
is always enabled on 81xx/83xx and HW does insert these fields into
CQE_RX. As a result receive buffer addresses will now be present at
7th word of CQE_RX instead of 6th.
Enable CQE_RX2_S on 88xx pass 2.x as
From: Sunil Goutham
This patch adds info on HW maximums of 81xx/83xx and also
configures receive and transmit datapaths accordingly.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c| 87 ++-
drivers/net/ethernet/cavium/thunder/nic_reg.h
From: Radoslaw Biernacki
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++--
drivers/net/ethernet/cavium/thunder/nicvf_m
From: Sunil Goutham
Set MAC addresses only for primary VF's and don't for
secondary VFs.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 2 +-
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --g
From: Sunil Goutham
This patch adds support for BGX module on 81xx where a BGX
can be split and have different LMACs configured in
different modes.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 112 --
drivers/net/ethernet/cavium/thund
From: Jerin Jacob
When SQ/TXQ is reclaimed i.e reset it's stats also automatically reset
by HW. This is not the case with RQ. Also VF doesn't have write access
to statistics counter registers. Hence a new Mbox msg is introduced which
supports resetting RQ, SQ and full Qset stats. Currently only R
From: Sunil Goutham
Instead of a round about way of converting buffers to SKBs and
combining them into a frag list, use standard skb_add_rx_frag()
API to merge page fragments. This code is useful when incoming
packets are of size more than RCV_FRAG_LEN which is currently
set to 2048bytes.
Signed
From: Sunil Goutham
On 88xx all LMACs in a BGX will be in same mode but on 81xx
BGX can be split as two and there can be LMACs configured in
different modes.
These changes move lmac_type, lane2serdes fields into per lmac
struct from BGX struct. Got rid of qlm_mode field which has become
redundan
From: Sunil Goutham
This patch adds RGX/RGMII interface type support to BGX
driver. This type of interface is supported by 81xx SOC.
CN81XX VNIC has 8 VFs and max possible LMAC interfaces are 9,
hence RGMII interface will not work if all DLMs are in BGX mode
and all 8 LMACs are enabled.
Signed-
From: Sunil Goutham
This patch series adds support for VNIC on 81xx and 83xx SOCs.
81xx/83xx is different from 88xx in terms of capabilities and different
types of interfaces supported (eg: QSGMII, RGMII) and have DLMs instead
of QLMs which allows single BGX to have interfaces of different LMAC
From: Sunil Goutham
This patch adds PCI device IDs of VNIC on newer chips and also
registers VF driver with them. Device id remains same for all
versions of chips but subsystem device id changes.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h| 10 +-
From: Sunil Goutham
TL4 calculation for a given SQ of secondary Qsets is incorrect
and goes out of bounds and also for some SQ's TL4 chosen will
transmit data via a different BGX interface and not same as
primary Qset's interface.
This patch fixes this issue.
Signed-off-by: Sunil Goutham
---
From: Sunil Goutham
Check for SMU RX local/remote faults along with SPU LINK
status. Otherwise at times link is UP at our end but DOWN
at link partner's side. Also due to an issue in BGX it's
rarely seen that initialization doesn't happen properly
and SMU RX reports faults with everything fine at
From: Sunil Goutham
This 2 patch series fixes issues w.r.t physical link status
reporting and transmit datapath configuration for
secondary qsets.
Changes from v1:
Fixed lmac disable sequence for interfaces of type SGMII.
Sunil Goutham (2):
net: thunderx: Fix link status reporting
net: thun
From: Sunil Goutham
Check for SMU RX local/remote faults along with SPU LINK
status. Otherwise at times link is UP at our end but DOWN
at link partner's side. Also due to an issue in BGX it's
rarely seen that initialization doesn't happen properly
and SMU RX reports faults with everything fine at
From: Sunil Goutham
TL4 calculation for a given SQ of secondary Qsets is incorrect
and goes out of bounds and also for some SQ's TL4 chosen will
transmit data via a different BGX interface and not same as
primary Qset's interface.
This patch fixes this issue.
Signed-off-by: Sunil Goutham
---
From: Sunil Goutham
This 2 patch series fixes issues w.r.t physical link status
reporting and transmit datapath configuration for
secondary qsets.
Sunil Goutham (2):
net: thunderx: Fix link status reporting
net: thunderx: Fix TL4 configuration for secondary Qsets
drivers/net/ethernet/cav
Thanks for fixing, changes look okay.
Sunil.
On Sun, May 8, 2016 at 3:46 PM, Heinrich Schuchardt wrote:
> Reserved fields should be set to zero to avoid exposing
> bits from the kernel stack.
>
> Signed-off-by: Heinrich Schuchardt
> ---
> drivers/net/ethernet/cavium/thunder/nicvf_queues.c | 4
From: Radoslaw Biernacki
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++--
drivers/net/ethernet/cavium/thunder/nicvf
From: Radoslaw Biernacki
With this sysfs attribute (sriov_sqs_assignment) administrator will be
able to read the current assigment of SQS/SVF for a given VF. This is
useful to decide which VFs needs to be attached to UIO for a successful
allocation of secondary Qsets
Signed-off-by: Radoslaw Bier
From: Radoslaw Biernacki
This patch adds support to PF for allocating additional Qsets to
dataplane apps such as DPDK. Till now PF, upon host bound interface's
request it used to allocate Qsets from the free ones, but for dataplane
apps support has been added for it to request specific Qsets inst
From: Jerin Jacob
Write access to VF statistics counter register is only allowed from PF.
Added a new mailbox message to reset VF's Rx/Tx counters, this is used
by userspace DPDK.
Signed-off-by: Jerin Jacob
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h | 27
From: Sunil Goutham
This patch series mainly adds support for userspace application
like DPDK with a VNIC VF attached to request additional QSets
for having morethan the default 8 queues.
Changes from v1:
Fixed compilation issue reported by kbuild test robot due to changes
in 2nd patch. Now 'nic
From: Radoslaw Biernacki
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++--
drivers/net/ethernet/cavium/thunder/nicvf
From: Radoslaw Biernacki
With this sysfs attribute (sriov_sqs_assignment) administrator will be
able to read the current assigment of SQS/SVF for a given VF. This is
useful to decide which VFs needs to be attached to UIO for a successful
allocation of secondary Qsets
Signed-off-by: Radoslaw Bier
From: Radoslaw Biernacki
This patch adds support to PF for allocating additional Qsets to
dataplane apps such as DPDK. Till now PF, upon host bound interface's
request it used to allocate Qsets from the free ones, but for dataplane
apps support has been added for it to request specific Qsets inst
From: Sunil Goutham
This patch series mainly adds support for userspace application
like DPDK with a VNIC VF attached to request additional QSets
for having morethan the default 8 queues.
Jerin Jacob (1):
net: thunderx: Introduce a mailbox message to reset VF counters
Radoslaw Biernacki (3):
From: Jerin Jacob
Write access to VF statistics counter register is only allowed from PF.
Added a new mailbox message to reset VF's Rx/Tx counters, this is used
by userspace DPDK.
Signed-off-by: Jerin Jacob
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h | 27
From: Sunil Goutham
Adjusted nicvf structure such that all elements used in hot
path like napi, xmit e.t.c fall into same cache line. This reduced
no of cache misses and resulted in ~2% increase in no of packets
handled on a core.
Also modified elements with :1 notation to boolean, to be
consist
From: Sunil Goutham
Instead of calling get_page() for every receive buffer carved out
of page, set page's usage count at the end, to reduce no of atomic
calls.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h |1 +
drivers/net/ethernet/cavium/thunder/nicv
From: Sunil Goutham
Below patches attempts to improve performance by reducing
no of atomic operations while allocating new receive buffers
and reducing cache misses by adjusting nicvf structure elements.
Changes from v1:
No changes, resubmitting a fresh as per David's suggestion.
Sunil Goutham
On Fri, Mar 11, 2016 at 1:18 AM, David Miller wrote:
> From: Sunil Kovvuri
> Date: Thu, 10 Mar 2016 23:57:48 +0530
>
>> Difference between NIU driver and this patch is there it's
>> calculate split count, increment page count and then divide page into
>> buf
>
> So calculate the modulus on the page split count and optimize the
> increment ahead of time when possible, and for the sub page split
> pieces do it one at a time.
>
Patch does almost the same with a negligible overhead of a counter
for page->_count increment at a later time but still before HW
Hi David,
>> So if you know ahead of time how the page will be split up, just
>> calculate that when you get the page and increment the page count
>> appropriately.
>>
>> That's what we do in the NIU driver.
>
> Thanks for the suggestion, will check and get back.
>
I looked at the NIU driver and
On Mon, Mar 7, 2016 at 10:34 PM, David Miller wrote:
> From: Sunil Kovvuri
> Date: Mon, 7 Mar 2016 22:28:39 +0530
>
>> Hi David,
>>
>>>> you create a window of time during which the consumer
>>>> can release the page and prematurely free it.
>>
Hi David,
>> you create a window of time during which the consumer
>> can release the page and prematurely free it.
Okay, but here the consumer i.e HW is notified only after page count
is incremented.
For example if you check 'nicvf_refill_rbdr' fn() only after receive
buffer ring is
refilled with
From: Sunil Goutham
Adjusted nicvf structure such that all elements used in hot
path like napi, xmit e.t.c fall into same cache line. This reduced
no of cache misses and resulted in ~2% increase in no of packets
handled on a core.
Also modified elements with :1 notation to boolean, to be
consist
From: Sunil Goutham
Instead of calling get_page() for every receive buffer carved out
of page, set page's usage count at the end, to reduce no of atomic
calls.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nic.h |1 +
drivers/net/ethernet/cavium/thunder/nicv
From: Sunil Goutham
Below patches attempts to improve performance by reducing
no of atomic operations while allocating new receive buffers
and reducing cache misses by adjusting nicvf structure elements.
Sunil Goutham (2):
net: thunderx: Set recevie buffer page usage count in bulk
net: thund
From: Sunil Goutham
On Thunderx pass 1.x and pass2 due to a HW errata default CQ
DROP_LEVEL of 0x80 is not sufficient to avoid CQ_WR_FULL Qset
error when packets are being received at >20Mpps resulting in
complete stall of packet reception.
This patch will configure it to 0x100 which is what is
From: Sunil Goutham
For secondary Qsets 'hw_tso' is not getting set as probe() returns
much earlier. Fixed it by moving silicon revision check.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nicvf_main.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
dif
From: Sunil Goutham
This patch series fixes couple of issues w.r.t multiqset mode
and receive packet statastics.
Sunil Goutham (3):
net: thunderx: Fix for multiqset not configured upon interface toggle
net: thunderx: Fix for HW TSO not enabled for secondary qsets
net: thunderx: Fix receive
From: Sunil Goutham
Counting rx packets for every CQE_RX in CQ irq handler is incorrect.
Synchronization is missing when multiple queues are receiving packets
simultaneously. Like transmit packet stats use HW stats here.
Also removed unused 'cqe_type' parameter in nicvf_rcv_pkt_handler().
Signe
From: Sunil Goutham
When a interface is assigned morethan 8 queues and the logical interface
is toggled i.e down & up, additional queues or qsets are not initialized
as secondary qset count is being set to zero while tearing down.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/th
From: Thanneeru Srinivasulu
When system is low on atomic memory, too many error messages are logged.
Since this is not a total failure but a simple switch to non-atomic allocation
better to have a stat.
Also add a stat for reset, kicked due to transmit watchdog timeout.
Signed-off-by: Thanneeru
From: Sunil Goutham
Allocate higher order pages when pagesize is small, this will
reduce number of calls to page allocator and wastage of memory.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nicvf_queues.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
dif
From: David Daney
In the case of OF device tree, the firmware information is attached to
the BGX device structure in the standard manner, so use the firmware
iterators and accessors where possible.
Signed-off-by: David Daney
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder
From: Sunil Goutham
This affinity hint can be used by user space irqbalance tool to set
preferred CPU mask for irqs registered by this VF. Irqbalance needs
to be in 'exact' mode to set irq affinity same as indicated by
affinity hint.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium
From: Robert Richter
Signed-off-by: Robert Richter
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 11 ---
1 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
b/drivers/net/ethernet/cav
From: Sunil Goutham
napi_schedule is being called from hard irq context, hence
switch to napi_schedule_irqoff which avoids unneeded call
to local_irq_save and local_irq_restore.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nicvf_main.c |2 +-
1 files changed, 1 inse
From: Sunil Goutham
This patch series contains changes
- To add support for virtual function's irq affinity hint
- Replace napi_schedule() with napi_schedule_irqoff()
- Reduce page allocation overhead by allocating pages
of higher order when pagesize is 4KB.
- Add couple of stats which helps in
Hi David,
If time permits, can you please look at this patchset.
Thanks,
Sunil.
On Mon, Feb 8, 2016 at 5:37 PM, wrote:
> From: Sunil Goutham
>
> This patch series contains changes
> - To add support for virtual function's irq affinity hint
> - Replace napi_schedule() with napi_schedule_irqoff
From: David Daney
In the case of OF device tree, the firmware information is attached to
the BGX device structure in the standard manner, so use the firmware
iterators and accessors where possible.
Signed-off-by: David Daney
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder
From: Sunil Goutham
This affinity hint can be used by user space irqbalance tool to set
preferred CPU mask for irqs registered by this VF. Irqbalance needs
to be in 'exact' mode to set irq affinity same as indicated by
affinity hint.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium
From: Sunil Goutham
Allocate higher order pages when pagesize is small, this will
reduce number of calls to page allocator and wastage of memory.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nicvf_queues.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
dif
From: Thanneeru Srinivasulu
When system is low on atomic memory, too many error messages are logged.
Since this is not a total failure but a simple switch to non-atomic allocation
better to have a stat.
Also add a stat for reset, kicked due to transmit watchdog timeout.
Signed-off-by: Thanneeru
From: Robert Richter
Signed-off-by: Robert Richter
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 11 ---
1 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
b/drivers/net/ethernet/cav
From: Sunil Goutham
napi_schedule is being called from hard irq context, hence
switch to napi_schedule_irqoff which avoids unneeded call
to local_irq_save and local_irq_restore.
Signed-off-by: Sunil Goutham
---
drivers/net/ethernet/cavium/thunder/nicvf_main.c |2 +-
1 files changed, 1 inse
From: Sunil Goutham
This patch series contains changes
- To add support for virtual function's irq affinity hint
- Replace napi_schedule() with napi_schedule_irqoff()
- Reduce page allocation overhead by allocating pages
of higher order when pagesize is 4KB.
- Add couple of stats which helps in
Will take care of above suggestions and resubmit.
Thanks,
Sunil.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tu
>> + booltns_mode:1;
>> + boolsqs_mode:1;
>These little refactors are creeping in your code without even being
mentioned in the commit message, this is not good practice.
Okay, will include these in the commit message.
>IMHO. Additionally, may be t
>
> If the performance increase is 4 %, then surely using twice more memory
> is not worth it.
I haven't mentioned anywhere that i am seeing 4% increase in performance.
Anyways as said already i will recheck and resubmit.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>After getting it working in guest i tried to apply it to host. With total of
>128 virtual functions (= 128 interfaces) it does not work at all.
> Even after bumping cma region size to insane value of 2GB more than half of
> interfaces still failed to allocate queues.
> And after setting cma=3G i
Thanks for the confirmation.
On Wed, Dec 2, 2015 at 5:00 PM, Pavel Fedin wrote:
> Hello again!
>
>> Subject: RE: [PATCH v2 4/5] net: thunderx: Switchon carrier only upon
>> interface link up
>>
>> Just a reminder, we have issue with this one too, which is not addressed
>> yet.
>
> I have exa
ill see the same issue.
And above i have suggested what could be done to not see this issue.
But anyway for the time being I will resubmit the patch series without this
patch, hope that would be okay.
On Wed, Dec 2, 2015 at 1:00 AM, David Miller wrote:
> From: Sunil Kovvuri
> Date: Tue,
Hi Pavel Fedin,
Are running Fedora 21 on Cavium ThunderX ?
Do you see linkup notification (dmesg) ?
If you see the existing driver (pasted snippet below), it does call
netif_carrier_on() upon receiving l
ink up notification from BGX driver.
===
case NIC_MBOX_MSG_BGX_LINK_CHANGE:
201 - 300 of 316 matches
Mail list logo
