Allocate the FIFO across the hardware Rx queues based on the priority
of the queues. Giving more FIFO resources to queues with a higher
priority. If PFC is active but not enabled for a queue, then less
resources can allocated to the queue.
Signed-off-by: Tom Lendacky
Add support to be able to use clause 37 auto-negotiation.
Signed-off-by: Tom Lendacky
---
drivers/net/ethernet/amd/xgbe/xgbe-common.h | 41 +
drivers/net/ethernet/amd/xgbe/xgbe-mdio.c | 242 ++-
drivers/net/ethernet/amd/xgbe/xgbe.h
Prepare the code to be able to work with more than one type of phy by
adding additional callable functions into the phy interface and removing
phy specific settings/functions from non-phy related files.
Signed-off-by: Tom Lendacky
---
Tx and Rx DMA channel status determiniation is different depending on the
version of the hardware. Update the channel status processing code to
account for the change. Also, reduce the timeout value used when stopping
the channels.
Signed-off-by: Tom Lendacky
---
Prepare the code to be able to support accessing of the PCS registers
in a new way, while maintaining the current access method. Provide a
version specific field that indicates the method to use.
Signed-off-by: Tom Lendacky
---
Currently, the Rx and Tx fifos are evenly allocated between the hardware
queues of the device. As more queues are instantiated, the fifo memory
needs to be able to be allocated based on queue priority. This allows for
higher priority queues to have more fifo memory than lower priority
queues.
Fix the length value used for the PCS register dump so that the full
value can be displayed.
Signed-off-by: Tom Lendacky
---
drivers/net/ethernet/amd/xgbe/xgbe-mdio.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git
This patch series is targeted at preparing the driver for a new PCI version
of the hardware. After this series is applied, a follow-on series will
introduce the support for the PCI version of the hardware.
The following updates and fixes are included in this driver update series:
- Fix
On Thu, Nov 3, 2016 at 6:52 PM, Marcelo Ricardo Leitner
wrote:
> On Thu, Nov 03, 2016 at 06:11:01PM +0100, Andrey Konovalov wrote:
>> On Wed, Nov 2, 2016 at 11:42 PM, Andrey Konovalov
>> wrote:
>> > On Wed, Oct 19, 2016 at 6:57 PM, Marcelo
This is the second time i am sending you this mail.I, Friedrich Mayrhofer
Donate $ 1,000,000.00 to You, Email Me personally for more details.
Regards.
Friedrich Mayrhofer
On Thu, Nov 03, 2016 at 06:11:01PM +0100, Andrey Konovalov wrote:
> On Wed, Nov 2, 2016 at 11:42 PM, Andrey Konovalov
> wrote:
> > On Wed, Oct 19, 2016 at 6:57 PM, Marcelo Ricardo Leitner
> > wrote:
> >> On Wed, Oct 19, 2016 at 02:25:24PM +0200,
ip_route_me_harder is not considering the L3 domain and sending lookups
to the wrong table. For example consider the following output rule:
iptables -I OUTPUT -p tcp --dport 12345 -j REJECT --reject-with tcp-reset
using perf to analyze lookups via the fib_table_lookup tracepoint shows:
vrf-test
Make sure to drop the references taken by of_get_child_by_name() and
bus_find_device() before returning from cpsw_phy_sel().
Note that holding a reference to the cpsw-phy-sel device does not
prevent the devres-managed private data from going away.
Fixes: 5892cd135e16 ("drivers: net:
Make sure to drop the reference taken by class_find_device() in
hnae_get_handle() on errors and when later releasing the handle.
Fixes: 6fe6611ff275 ("net: add Hisilicon Network Subsystem...")
Cc: Yisen Zhuang
Cc: Salil Mehta
Signed-off-by: Johan
Make sure to drop the reference taken by bus_find_device_by_name()
before returning from phy_connect() and phy_attach().
Note that both function still take a reference to the phy device
through phy_attach_direct().
Fixes: e13934563db0 ("[PATCH] PHY Layer fixup")
Cc: Florian Fainelli
This series fixes a number of device reference leaks (and one of_node
leak) due to failure to drop the references taken by bus_find_device()
and friends.
Note that the final two patches have been compile tested only.
Thanks,
Johan
v2
- hold reference to cpsw-phy-sel device while accessing
Make sure to drop the references taken by bus_find_device() before
returning from emac_dev_open().
Note that phy_connect still takes a reference to the phy device.
Fixes: 5d69e0076a72 ("net: davinci_emac: switch to new mdio")
Cc: Mugunthan V N
Cc: Grygorii Strashko
On Thu, Nov 3, 2016 at 2:18 AM, Eric Dumazet wrote:
> Lorenzo, have'nt you already fixed all these bugs ?
Not yet. There's still a fair bit of out-of-tree code left. Other than
per-UID routing, xt_qtaguid is the big one, but there's also xt_quota2
and xt_idletimer. to
- Use the UID in routing lookups made by protocol connect() and
sendmsg() functions.
- Make sure that routing lookups triggered by incoming packets
(e.g., Path MTU discovery) take the UID of the socket into
account.
- For packets not associated with a userspace socket, (e.g., ping
replies)
Protocol sockets (struct sock) don't have UIDs, but most of the
time, they map 1:1 to userspace sockets (struct socket) which do.
Various operations such as the iptables xt_owner match need
access to the "UID of a socket", and do so by following the
backpointer to the struct socket. This involves
This patchset adds support for per-UID routing. It allows the
administrator to configure rules such as:
ip rule add uidrange 100-200 lookup 123
This functionality has been in use by all Android devices since
5.0. It is primarily used to impose per-app routing policies (on
Android, every app
On 11/03/2016 12:14 PM, David Miller wrote:
>
> 20+ patches is too many to submit at one time.
>
> Making huge patch series submissions puts an unreasonable burdon on
> those developers who might decide to review your work, including me.
>
> Please keep your series down to a small, reasonable,
- Define a new FIB rule attributes, FRA_UID_RANGE, to describe a
range of UIDs.
- Define a RTA_UID attribute for per-UID route lookups and dumps.
- Support passing these attributes to and from userspace via
rtnetlink. The value INVALID_UID indicates no UID was
specified.
- Add a UID field to
Hi,
Another report that looks related:
[ INFO: possible circular locking dependency detected ]
4.9.0-rc3+ #344 Not tainted
---
syz-executor/25526 is trying to acquire lock:
([ 950.351060] [i].mutex
[] nfnl_lock+0x28/0x30
Ring allocation helpers encapsulate all ring allocation and
initialization steps nicely. Reuse them on .ndo_open() path.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 60 --
1 file changed, 20
From: Gao Feng
Date: Thu, 3 Nov 2016 22:38:28 +0800
> On Thu, Nov 3, 2016 at 10:31 PM, Eric Dumazet wrote:
>> On Thu, 2016-11-03 at 21:39 +0800, Gao Feng wrote:
>>> Hi Eric,
>>>
>>> On Thu, Nov 3, 2016 at 9:30 PM, Eric Dumazet
"Shadow" in ring helpers used to mean that the helper will allocate
rings without touching existing configuration, this was used for
reconfiguration while the device was running. We will soon use
the same helpers for .ndo_open() path, so replace "shadow" with
"ring_set".
No functional changes.
Upcoming XDP support will break the assumption that one can iterate
over IRQ vectors to get to all the rings easily. Use nn->.x_ring
arrays directly.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 12 ++--
1 file
20+ patches is too many to submit at one time.
Making huge patch series submissions puts an unreasonable burdon on
those developers who might decide to review your work, including me.
Please keep your series down to a small, reasonable, size. Perhaps
10-15 patches maximum.
I will not be
Calculate packet offsets early in nfp_net_rx() so that we will be
able to use them in upcoming XDP handler. While at it move relevant
variables into the loop scope.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 56
icmp_send is called in response to some event. The skb may not have
the device set (skb->dev is NULL), but it is expected to have an rt.
Update icmp_route_lookup to use the rt on the skb to determine L3
domain.
Fixes: 613d09b30f8b ("net: Use VRF device index for lookups on TX")
Signed-off-by:
Allow changing the number of rings via ethtool .set_channels API.
Runtime reconfig needs to be extended to handle number of rings.
We need to be able to activate interrupt vectors before rings are
assigned to them.
Signed-off-by: Jakub Kicinski
---
We will need to rerun the initialization of the RSS indirection table
after the number of rings is changed. Move the code to a separate
function.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 15 ++-
1 file
Report number of rings via ethtool .get_channels API.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c | 16
1 file changed, 16 insertions(+)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
Most infrastructure can be reused, provide separate handling
of context offsets and exit codes.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_bpf.h | 1 +
drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c | 92
Add XDP support. Separate stack's and XDP's TX rings logically.
Add functions for handling XDP_TX and cleanup of XDP's TX rings.
For XDP allocate all RX buffers as separate pages and map them
with DMA_BIDIRECTIONAL.
Signed-off-by: Jakub Kicinski
---
seq_file users can only access const version of file pointer,
because the ->file member of struct seq_operations is marked
as such. Make parameter to debugfs_real_fops() const.
CC: Greg Kroah-Hartman
CC: Nicolai Stange
CC: Christian Lamparter
nfp_net_bpf_offload() takes all .setup_tc() parameters but it
doesn't use them at the moment. Remove unnecessary ones to make
it possible for XDP to reuse this function.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net.h | 4
Instead of fixing ring -> vector relations up in ring swap functions
put the reassignment into a helper function which will reinit all
links.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 40 +-
1 file
Hi!
This set adds support for ethtool channel API and XDP.
I kick off with ethtool get_channels() implementation.
set_channels() needs some preparations to get right. I follow
the prepare/commit paradigm and allocate all resources before
stopping the device. It has already been done for
All functions which need to reallocate ring resources at runtime
look very similar. Centralize that logic into a separate function.
Encapsulate configuration parameters in a structure.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net.h
On Wed, Nov 2, 2016 at 11:42 PM, Andrey Konovalov wrote:
> On Wed, Oct 19, 2016 at 6:57 PM, Marcelo Ricardo Leitner
> wrote:
>> On Wed, Oct 19, 2016 at 02:25:24PM +0200, Andrey Konovalov wrote:
>>> Hi,
>>>
>>> I've got the following error report
And besides, one can argue that in the SR-IOV scenario the VF has no business
knowing the physical port speed.
Good point, but there are more use-cases we should consider.
For example, when using Multi-Host/Flex-10/Multi-PF each PF should
be able to query both physical port speed and actual
On 02/11/2016 17:50, Mintz, Yuval wrote:
>> Sending RFC to get feedback for the following ethtool proposal:
>>
>> In some cases such as virtual machines and multi functions (SR-IOV), the
>> actual
>> bandwidth exposed for each machine is not accurately shown in ethtool.
>> Currently ethtool
On Mon, 2016-10-31 at 11:25 +0100, André Roth wrote:
> Hi all,
>
> >
> > on my device this results in:
> > [0xc9410018] = 0x200
> > [0xc9410030] = 0x0
> > [0xc941003c] = 0x0
> > [0xc9411000] = 0x1100802
> > [0xc9411018] = 0x2202006
> > [0xc9411028] = 0x0
> >
> > maybe someone else could
On Thu, 03 Nov 2016 17:55:39 +0100, Nicolai Stange wrote:
> Hi Jakub,
>
> thanks for this.
>
> However, the debugfs maintainer, Greg K-H, as well as the lkml is
> missing from the To/Cc. Can you resend please?
Sure thing!
> Jakub Kicinski writes:
>
> > seq_file
Hi Jakub,
thanks for this.
However, the debugfs maintainer, Greg K-H, as well as the lkml is
missing from the To/Cc. Can you resend please?
Jakub Kicinski writes:
> seq_file users can only access const version of file pointer,
... because the ->file member of
On Thu, Nov 03, 2016 at 09:31:21AM +0100, Rafal Milecki wrote:
> On 11/02/2016 06:08 PM, Jon Mason wrote:
> >Change the bgmac driver to allow for phy's defined by the device tree
>
> This is a late review, I know, sorry... :(
>
>
> >+static int bcma_phy_direct_connect(struct bgmac *bgmac)
> >+{
cgroupstats_cmd_get_policy is [CGROUPSTATS_CMD_ATTR_MAX+1],
taskstats_cmd_get_policy[TASKSTATS_CMD_ATTR_MAX+1],
but their family.maxattr is TASKSTATS_CMD_ATTR_MAX.
CGROUPSTATS_CMD_ATTR_MAX is less than TASKSTATS_CMD_ATTR_MAX,
so we could end up accessing out-of-bound.
Change
In __genl_register_family(), when genl_validate_assign_mc_groups()
fails, we forget to free the memory we possibly allocate for
family->attrbuf.
Note, some callers call genl_unregister_family() to clean up
on error path, it doesn't work because the family is inserted
to the global list in the
On 11/03/2016 03:15 PM, Dmitry Vyukov wrote:
On Wed, Nov 2, 2016 at 11:14 PM, Dmitry Vyukov wrote:
Here we go.
The following program triggers kernel BUG in htab_elem_free.
On commit 0c183d92b20b5c84ca655b45ef57b3318b83eb9e (Oct 31).
Run as "while true; do ./a.out; done".
On Sat, 2016-10-01 at 17:58 +0200, Martin Blumenstingl wrote:
> Hello Peppe,
>
> On Mon, Sep 26, 2016 at 8:17 AM, Giuseppe CAVALLARO
> wrote:
> >
> > Hello André
> >
> > On 9/17/2016 11:23 PM, André Roth wrote:
> > >
> > >
> > >
> > > Hi all,
> > >
> > > I have an
Hi,
We experience some limit on the maximum packets in flight which seem not to be
related with the receive or write buffers. Does somebody know if there is an
issue with a maximum of around 1MByte (or sometimes 2Mbyte) of data in flight
per TCP flow?
It seems to be a strict and stable limit
Binding a raw socket to a local address fails if the socket is bound
to an L3 domain:
$ vrf-test -s -l 10.100.1.2 -R -I red
error binding socket: 99: Cannot assign requested address
Update raw_bind to look consider if sk_bound_dev_if is bound to an L3
domain and use inet_addr_type_table
Hello,
I'm seeing pretty strange issue with GMAC reporting a lot of link state changes
based on bits in GMAC_RGSMIIIS. It looks like that:
-->8---
Link is Down
Link is Up - 10/Full
Link is Down
Link is Up - 10/Half
Link is Down
Link is Down
Link is Up -
Em Thu, Nov 03, 2016 at 08:59:46AM -0700, Eric Dumazet escreveu:
> From: Eric Dumazet
>
> While fuzzing kernel with syzkaller, Andrey reported a nasty crash
> in inet6_bind() caused by DCCP lacking a required method.
Ouch, thanks, forgot the mapped case :-)
Acked-by:
From: Eric Dumazet
While fuzzing kernel with syzkaller, Andrey reported a nasty crash
in inet6_bind() caused by DCCP lacking a required method.
Fixes: ab1e0a13d7029 ("[SOCK] proto: Add hashinfo member to struct proto")
Signed-off-by: Eric Dumazet
On 11/3/2016 2:09 AM, Ruinskiy, Dima wrote:
-Original Message-
From: Intel-wired-lan [mailto:intel-wired-lan-boun...@lists.osuosl.org] On
Behalf Of Tyler Baicar
Sent: Wednesday, 02 November, 2016 23:08
To: Kirsher, Jeffrey T; intel-wired-...@lists.osuosl.org;
netdev@vger.kernel.org;
--
HELLO,
HAVE YOU RECEIVED THE FUNDS? PLEASE CONFIRM TO ME IF YOU HAVE RECEIVED
THE FUND OR NOT.
REV.JOHN GAZDA
Hi Eric,
It seems that your patch fixes the issue, I'm not seeing the report any more.
Tested-by: Andrey Konovalov
Thanks!
On Thu, Nov 3, 2016 at 4:39 PM, Eric Dumazet wrote:
> On Wed, Nov 2, 2016 at 2:14 PM, Andrey Konovalov
Hi,
Sorry for not chipping in earlier - LPC is taking my time.
> > > > Looks like we are missing a kfree(family->attrbuf); on error
> > > > path, but it is not related to Johannes' recent patches.
Actually, I think it *is* related to my patch - I inserted the code
there in the wrong place or
On 11/03/2016 03:41 AM, Joe Perches wrote:
On Sun, 2016-10-30 at 19:02 -0400, Jes Sorensen wrote:
Code is 80 characters wide, and comments are /* */ never the ugly C++
crap.
You might look at the recent Linus Torvalds authored commit
5e467652ffef (?printk: re-organize log_output() to be more
On Wed, Nov 2, 2016 at 2:14 PM, Andrey Konovalov wrote:
> Hi,
>
> I've got the following error report while running the syzkaller fuzzer:
>
> BUG: unable to handle kernel NULL pointer dereference at (null)
> IP: [< (null)>] (null)
> PGD 66b6f067
On 11/03/2016 02:10 AM, John Heenan wrote:
On 3 November 2016 at 11:00, Larry Finger wrote:
On 10/30/2016 05:20 AM, John Heenan wrote:
This fix enables the same sequence of init behaviour as the alternative
working driver for the wireless rtl8723bu IC at
Hi Cong,
Yes, the last patch fixes the issue.
Tested-by: Andrey Konovalov
Thanks!
On Thu, Nov 3, 2016 at 6:29 AM, Cong Wang wrote:
> On Wed, Nov 2, 2016 at 10:25 PM, Cong Wang wrote:
>> On Wed, Nov 2, 2016 at 5:25
On 3 November 2016 at 08:52, Hannes Frederic Sowa
wrote:
> On 02.11.2016 23:54, Thomas Graf wrote:
>> Why would I want to accept the overhead if I simply avoid it? Just
>> parsing the header and doing the hash lookup will add cost, cost for
>> each packet.
>
> That is
> There are a couple of things about this. Russel's work isn't part of
> the kernel yet so I can't make use of it.
Well, you could guide it into the kernel. Part of it has already made
the way in. And i know of other platforms which would benefit from it.
> Additionally, the I2C device is
Hi Eric,
On Thu, Nov 3, 2016 at 11:07 PM, Eric Dumazet wrote:
> On Thu, 2016-11-03 at 22:38 +0800, Gao Feng wrote:
>
>> Because other net devices put the statistics together.
>> Take tun/tap as example, it is a virtual device, but its all
>> statistics are percpu
On Thu, 2016-11-03 at 22:38 +0800, Gao Feng wrote:
> Because other net devices put the statistics together.
> Take tun/tap as example, it is a virtual device, but its all
> statistics are percpu including dropped.
Take a look at 2681128f0ced8aa4e66f221197e183cc16d244fe
("veth: reduce stat
> -Original Message-
> From: Christoph Hellwig [mailto:h...@infradead.org]
> Sent: Thursday, November 3, 2016 7:46 AM
> To: Konrad Rzeszutek Wilk
> Cc: Christoph Hellwig ; Duyck, Alexander H
> ; linux...@kvack.org;
On 11/03/2016 09:28 AM, Andrew Lunn wrote:
> On Thu, Nov 03, 2016 at 08:30:36AM -0500, Tom Lendacky wrote:
>> Add support to initialize and use the I2C controller within the hardware
>> in order to determine the SFP media type that is installed.
>
> Hi Tom
Hi Andrew
>
> Did you see the work
Install the callbacks via the state machine.
Cc: "David S. Miller"
Cc: netdev@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior
Signed-off-by: Thomas Gleixner
---
include/linux/cpuhotplug.h | 1 +
net/core/dev.c
On 11/03/2016 08:51 AM, Mintz, Yuval wrote:
>> +static void xgbe_calculate_equal_fifo(unsigned int fifo_size,
>> + unsigned int queue_count,
>> + unsigned int *fifo)
>> +{
> ...
>> +
>> +return;
>> }
>
> No need for explicit
On 02.11.2016 23:54, Thomas Graf wrote:
> On 1 November 2016 at 16:12, Hannes Frederic Sowa
> wrote:
>> On 01.11.2016 21:59, Thomas Graf wrote:
Dumping and verifying which routes get used might actually already be
quite complex on its own. Thus my fear.
>>>
Install the callbacks via the state machine. Use multi state support to avoid
custom list handling for the multiple instances.
Cc: "David S. Miller"
Cc: Steffen Klassert
Cc: Herbert Xu
Cc: netdev@vger.kernel.org
On Thu, Nov 03, 2016 at 10:29:52AM -0400, Konrad Rzeszutek Wilk wrote:
> Somehow I thought you wanted to put them through your tree (which
> is why I acked them).
>
> I can take them and also the first couple of Alexander through
> my tree. Or if it makes it simpler - they can go through the -mm
On Thu, 2016-11-03 at 21:39 +0800, Gao Feng wrote:
> Hi Eric,
>
> On Thu, Nov 3, 2016 at 9:30 PM, Eric Dumazet wrote:
> > On Thu, 2016-11-03 at 21:03 +0800, f...@ikuai8.com wrote:
> >> From: Gao Feng
> >>
> >> The dropped count of veth is located in
On Thu, Nov 3, 2016 at 10:31 PM, Eric Dumazet wrote:
> On Thu, 2016-11-03 at 21:39 +0800, Gao Feng wrote:
>> Hi Eric,
>>
>> On Thu, Nov 3, 2016 at 9:30 PM, Eric Dumazet wrote:
>> > On Thu, 2016-11-03 at 21:03 +0800, f...@ikuai8.com wrote:
>> >>
On Thu, Nov 03, 2016 at 07:14:46AM -0700, Christoph Hellwig wrote:
> On Wed, Nov 02, 2016 at 07:12:31AM -0400, Alexander Duyck wrote:
> > There are no users for swiotlb_map_sg or swiotlb_unmap_sg so we might as
> > well just drop them.
>
> FYI, I sent the same patch already on Sep, 11 and Konrad
On Thu, Nov 03, 2016 at 08:30:36AM -0500, Tom Lendacky wrote:
> Add support to initialize and use the I2C controller within the hardware
> in order to determine the SFP media type that is installed.
Hi Tom
Did you see the work Russell King did for phylink?
https://lwn.net/Articles/667055/
On Wed, Nov 2, 2016 at 11:14 PM, Dmitry Vyukov wrote:
> Here we go.
>
> The following program triggers kernel BUG in htab_elem_free.
> On commit 0c183d92b20b5c84ca655b45ef57b3318b83eb9e (Oct 31).
> Run as "while true; do ./a.out; done".
>
> [ cut here ]
On 2 November 2016 at 04:48, Hannes Frederic Sowa
wrote:
> On Wed, Nov 2, 2016, at 00:07, Tom Herbert wrote:
>> On the other hand, I'm not really sure how to implement for this level
>> of performance this in LWT+BPF either. It seems like one way to do
>> that would be
> +static void xgbe_calculate_equal_fifo(unsigned int fifo_size,
> + unsigned int queue_count,
> + unsigned int *fifo)
> +{
...
> +
> + return;
> }
No need for explicit return.
On Wed, Nov 02, 2016 at 07:12:31AM -0400, Alexander Duyck wrote:
> There are no users for swiotlb_map_sg or swiotlb_unmap_sg so we might as
> well just drop them.
FYI, I sent the same patch already on Sep, 11 and Konrad already ACKed
it:
https://lkml.org/lkml/2016/9/11/112
Calculate packet offsets early in nfp_net_rx() so that we will be
able to use them in upcoming XDP handler. While at it move relevant
variables into the loop scope.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 56
"Shadow" in ring helpers used to mean that the helper will allocate
rings without touching existing configuration, this was used for
reconfiguration while the device was running. We will soon use
the same helpers for .ndo_open() path, so replace "shadow" with
"ring_set".
No functional changes.
This patch provides support for the presence of a KR redriver chip in
between the device PCS and an external PHY. When a redriver chip is
present the device must perform clause 73 auto-negotiation in order to
set the redriver chip for the downstream connection.
Signed-off-by: Tom Lendacky
Fix the length value used for the PCS register dump so that the full
value can be displayed.
Signed-off-by: Tom Lendacky
---
drivers/net/ethernet/amd/xgbe/xgbe-mdio.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git
Allow changing the number of rings via ethtool .set_channels API.
Runtime reconfig needs to be extended to handle number of rings.
We need to be able to activate interrupt vectors before rings are
assigned to them.
Signed-off-by: Jakub Kicinski
---
Most infrastructure can be reused, provide separate handling
of context offsets and exit codes.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_bpf.h | 1 +
drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c | 92
Instead of fixing ring -> vector relations up in ring swap functions
put the reassignment into a helper function which will reinit all
links.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 40 +-
1 file
The default TX queue length of Ethernet devices have been a magic
constant of 1000, ever since the initial git import.
Looking back in historical trees[1][2] the value used to be 100,
with the same comment "Ethernet wants good queues". The commit[3]
that changed this from 100 to 1000 didn't
We will need to rerun the initialization of the RSS indirection table
after the number of rings is changed. Move the code to a separate
function.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 15 ++-
1 file
All functions which need to reallocate ring resources at runtime
look very similar. Centralize that logic into a separate function.
Encapsulate configuration parameters in a structure.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net.h
Add XDP support. Separate stack's and XDP's TX rings logically.
Add functions for handling XDP_TX and cleanup of XDP's TX rings.
For XDP allocate all RX buffers as separate pages and map them
with DMA_BIDIRECTIONAL.
Signed-off-by: Jakub Kicinski
---
Ring allocation helpers encapsulate all ring allocation and
initialization steps nicely. Reuse them on .ndo_open() path.
Signed-off-by: Jakub Kicinski
---
.../net/ethernet/netronome/nfp/nfp_net_common.c| 60 --
1 file changed, 20
Upcoming XDP support will break the assumption that one can iterate
over IRQ vectors to get to all the rings easily. Use nn->.x_ring
arrays directly.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 12 ++--
1 file
seq_file users can only access const version of file pointer,
make parameter to debugfs_real_fops() const.
CC: Nicolai Stange
CC: Christian Lamparter
Signed-off-by: Jakub Kicinski
---
include/linux/debugfs.h | 3 ++-
1
nfp_net_bpf_offload() takes all .setup_tc() parameters but it
doesn't use them at the moment. Remove unnecessary ones to make
it possible for XDP to reuse this function.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net.h | 4
Report number of rings via ethtool .get_channels API.
Signed-off-by: Jakub Kicinski
---
drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c | 16
1 file changed, 16 insertions(+)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
101 - 200 of 255 matches
Mail list logo