Hi Catherine,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net-next/master]
url:
https://github.com/0day-ci/linux/commits/Catherine-Sullivan/Add-gve-driver/20190629-070444
reproduce:
# apt-get install sparse
# sparse version: v0.6.1-rc1-7-g2b96
Fixes: ac0744578517 ("gve: Add ethtool support")
Signed-off-by: kbuild test robot
---
gve_ethtool.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/google/gve/gve_ethtool.c
b/drivers/net/ethernet/google/gve/gve_ethtool.c
index 8e6863c..0363
On Thu, Jun 27, 2019 at 2:45 PM Andrii Nakryiko
wrote:
>
> On Thu, Jun 27, 2019 at 2:04 PM Daniel Borkmann wrote:
> >
> > On 06/26/2019 08:12 AM, Andrii Nakryiko wrote:
> > > BPF_MAP_TYPE_PERF_EVENT_ARRAY map is often used to send data from BPF
> > > program
> > > to user space for additional pr
For BPF_MAP_TYPE_PERF_EVENT_ARRAY typically correct size is number of
possible CPUs. This is impossible to specify at compilation time. This
change adds automatic setting of PERF_EVENT_ARRAY size to number of
system CPUs, unless non-zero size is specified explicitly. This allows
to adjust size for
Add test verifying perf buffer API functionality.
Signed-off-by: Andrii Nakryiko
Acked-by: Song Liu
---
.../selftests/bpf/prog_tests/perf_buffer.c| 94 +++
.../selftests/bpf/progs/test_perf_buffer.c| 29 ++
2 files changed, 123 insertions(+)
create mode 100644 tools
This patchset adds a high-level API for setting up and polling perf buffers
associated with BPF_MAP_TYPE_PERF_EVENT_ARRAY map. Details of APIs are
described in corresponding commit.
Patch #1 adds a set of APIs to set up and work with perf buffer.
Patch #2 enhances libbpf to support auto-setting PE
BPF_MAP_TYPE_PERF_EVENT_ARRAY map is often used to send data from BPF program
to user space for additional processing. libbpf already has very low-level API
to read single CPU perf buffer, bpf_perf_event_read_simple(), but it's hard to
use and requires a lot of code to set everything up. This patch
Switch event_pipe implementation to rely on new libbpf perf buffer API
(it's raw low-level variant).
Signed-off-by: Andrii Nakryiko
---
tools/bpf/bpftool/map_perf_ring.c | 207 ++
1 file changed, 68 insertions(+), 139 deletions(-)
diff --git a/tools/bpf/bpftool/map_p
This is a respin of https://www.spinics.net/lists/netdev/msg526066.html
Since page_pool API fixes are merged into net-next we can now safely use
it's DMA mapping capabilities.
First patch changes the buffer allocation from napi/netdev_alloc_frag()
to page_pool API. Although this will lead to sli
Use page_pool and it's DMA mapping capabilities for Rx buffers instead
of netdev/napi_alloc_frag()
Although this will result in a slight performance penalty on small sized
packets (~10%) the use of the API will allow to easily add XDP support.
The penalty won't be visible in network testing i.e ip
Since the dma direction is stored in page pool params, offer an API
helper for driver that choose not to keep track of it locally
Signed-off-by: Ilias Apalodimas
Acked-by: Jesper Dangaard Brouer
---
include/net/page_pool.h | 9 +
1 file changed, 9 insertions(+)
diff --git a/include/net
The interface only supports 1 Tx queue so locking is introduced on
the Tx queue if XDP is enabled to make sure .ndo_start_xmit and
.ndo_xdp_xmit won't corrupt Tx ring
- Performance (SMMU off)
Benchmark XDP_SKB XDP_DRV
xdp1291kpps 344kpps
rxdrop 282kpps 342kpps
- Perfor
Hi Catherine,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net-next/master]
url:
https://github.com/0day-ci/linux/commits/Catherine-Sullivan/Add-gve-driver/20190629-070444
reproduce:
# apt-get install sparse
# sparse version: v0.6.1-rc1-7-g2b96
Fixes: fa090987329c ("gve: Add transmit and receive support")
Signed-off-by: kbuild test robot
---
gve_main.c |6 +++---
gve_rx.c |2 +-
gve_tx.c |2 +-
3 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/google/gve/gve_main.c
b/drivers/net/ether
Use new bpf_program__attach_perf_event() in test previously relying on
direct ioctl manipulations.
Signed-off-by: Andrii Nakryiko
Reviewed-by: Stanislav Fomichev
Acked-by: Song Liu
---
.../bpf/prog_tests/stacktrace_build_id_nmi.c | 31 +--
1 file changed, 15 insertions(+), 16
Allow attaching BPF programs to kernel tracepoint BPF hooks specified by
category and name.
Signed-off-by: Andrii Nakryiko
Acked-by: Song Liu
---
tools/lib/bpf/libbpf.c | 79
tools/lib/bpf/libbpf.h | 4 ++
tools/lib/bpf/libbpf.map | 1 +
3 files ch
Convert some existing tests that attach to tracepoints to use
bpf_program__attach_tracepoint API instead.
Signed-off-by: Andrii Nakryiko
Reviewed-by: Stanislav Fomichev
Acked-by: Song Liu
---
.../bpf/prog_tests/stacktrace_build_id.c | 50 ---
.../selftests/bpf/prog_tests/s
Add a wrapper utilizing bpf_link "infrastructure" to allow attaching BPF
programs to raw tracepoints.
Signed-off-by: Andrii Nakryiko
Acked-by: Song Liu
---
tools/lib/bpf/libbpf.c | 37 +
tools/lib/bpf/libbpf.h | 3 +++
tools/lib/bpf/libbpf.map | 1 +
3
Add tests verifying kprobe/kretprobe/uprobe/uretprobe APIs work as
expected.
Signed-off-by: Andrii Nakryiko
Reviewed-by: Stanislav Fomichev
Acked-by: Song Liu
---
.../selftests/bpf/prog_tests/attach_probe.c | 155 ++
.../selftests/bpf/progs/test_attach_probe.c | 55 +++
bpf_link is and abstraction of an association of a BPF program and one
of many possible BPF attachment points (hooks). This allows to have
uniform interface for detaching BPF programs regardless of the nature of
link and how it was created. Details of creation and setting up of
a specific bpf_link
Add ability to attach to kernel and user probes and retprobes.
Implementation depends on perf event support for kprobes/uprobes.
Signed-off-by: Andrii Nakryiko
---
tools/lib/bpf/libbpf.c | 165 +++
tools/lib/bpf/libbpf.h | 7 ++
tools/lib/bpf/libbpf.map
This patchset adds the following APIs to allow attaching BPF programs to
tracing entities:
- bpf_program__attach_perf_event for attaching to any opened perf event FD,
allowing users full control;
- bpf_program__attach_kprobe for attaching to kernel probes (both entry and
return probes);
- bpf_p
bpf_program__attach_perf_event allows to attach BPF program to existing
perf event hook, providing most generic and most low-level way to attach BPF
programs. It returns struct bpf_link, which should be passed to
bpf_link__destroy to detach and free resources, associated with a link.
Signed-off-by
It's often inconvenient to switch sign of error when passing it into
libbpf_strerror_r. It's better for it to handle that automatically.
Signed-off-by: Andrii Nakryiko
Reviewed-by: Stanislav Fomichev
Acked-by: Song Liu
---
tools/lib/bpf/str_error.c | 2 +-
1 file changed, 1 insertion(+), 1 del
On Fri, 28 Jun 2019 17:59:25 -0700, Jakub Kicinski wrote:
> > > Sorry for all the questions, I'm not really able to fully wrap my head
> > > around this. I also feel like I'm missing the sockmap piece that may
> > > be why you prefer unhash over disconnect.
> >
> > Yep, if we try to support li
On 6/26/2019 11:52 AM, Catherine Sullivan wrote:
[snip]
> +static void
> +gve_get_ethtool_stats(struct net_device *netdev,
> + struct ethtool_stats *stats, u64 *data)
> +{
> + struct gve_priv *priv = netdev_priv(netdev);
> + u64 rx_pkts, rx_bytes, tx_pkts, tx_bytes;
>
Fixes: ac22601d5bb7 ("gve: Add basic driver framework for Compute Engine
Virtual NIC")
Signed-off-by: kbuild test robot
---
gve_main.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/google/gve/gve_main.c
b/drivers/net/ethernet/google/gve/gve_main.c
Hi Catherine,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net-next/master]
url:
https://github.com/0day-ci/linux/commits/Catherine-Sullivan/Add-gve-driver/20190629-070444
reproduce:
# apt-get install sparse
# sparse version: v0.6.1-rc1-7-g2b96
On Fri, 28 Jun 2019 17:20:23 -0700, John Fastabend wrote:
> > Why can't tls sockets exist outside of established state? If shutdown
> > doesn't call close, perhaps we can add a shutdown callback? It doesn't
> > seem to be called from BH?
> >
>
> Because the ulp would be shared in this case,
>
Jakub Kicinski wrote:
> On Fri, 28 Jun 2019 12:40:29 -0700, John Fastabend wrote:
> > The lock() is already held when entering unhash() side so need to
> > handle this case as well,
> >
> > CPU 0 (free) CPU 1 (wq)
> >
> > lock(sk) ctx = tls_get_ctx(sk) <- need to be check nu
On 06/28/2019 06:08 AM, Baruch Siach wrote:
> Merge commit 1c8c5a9d38f60 ("Merge
> git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next") undid the
> fix from commit 36f9814a494 ("bpf: fix uapi hole for 32 bit compat
> applications") by taking the gpl_compatible 1-bit field definition from
On 06/28/2019 11:12 AM, Toke Høiland-Jørgensen wrote:
> When using the bpf_redirect_map() helper to redirect packets from XDP, the
> eBPF
> program cannot currently know whether the redirect will succeed, which makes
> it
> impossible to gracefully handle errors. To properly fix this will probabl
On 06/28/2019 03:12 AM, Stanislav Fomichev wrote:
> Let's use union with u8[4] and u32 members for sockopt buffer,
> that should fix any possible aliasing issues.
>
> test_sockopt_sk.c: In function ‘getsetsockopt’:
> test_sockopt_sk.c:115:2: warning: dereferencing type-punned pointer will
> break
On 6/28/19 3:54 AM, Denis Kirjanov wrote:
> Current code assumes that we print ethernet mac and
> that doesn't work in the IPoIB case with SRIOV-enabled hardware
>
> Before:
> 11: ib1: mtu 2044 qdisc pfifo_fast
> state UP mode DEFAULT group default qlen 256
> link/infiniband
> 80:00:00:66
Hi Dave,
This series adds misc updates to mlx5e driver.
For more information please see tag log below.
Please pull and let me know if there is any problem.
Please note that the series starts with a merge of mlx5-next branch,
to resolve and avoid dependency with rdma tree.
This pull provides the
From: Gavi Teitz
Refreshing TIRs is done in order to update the TIRs with the current
state of SQs in the transport domain, so that the TIRs can filter out
undesired self-loopback packets based on the source SQ of the packet.
Representor TIRs will only receive packets that originate from their
a
From: Paul Blakey
After changing the parent_id to be the same for both NICs of same
the hardware device, netdev_port_same_parent_id now returns true for
more cases (all the lower devices in the hierarchy are on the same
hardware device).
If merged eswitch isn't enabled, these cases aren't suppor
From: Paul Blakey
Report system_image_guid as the E-Switch switch_id, this ensures
that when a NIC contains multiple PCI functions and which
has merged eswitch capability, all representors from
multiple PFs publish same switch_id.
Signed-off-by: Paul Blakey
Reviewed-by: Parav Pandit
Reviewed-b
From: Arnd Bergmann
Putting an empty 'mlx5_flow_spec' structure on the stack is a bit
wasteful and causes a warning on 32-bit architectures when building
with clang -fsanitize-coverage:
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c: In function
'mlx5_eswitch_termtbl_create'
From: Parav Pandit
Currently PF phys_port_name is named as pfNvf-1 as vport number for PF
vport is 65535.
Correct PF's phys_port name as agreed upon name as pfN.
Signed-off-by: Parav Pandit
Reviewed-by: Vu Pham
Signed-off-by: Saeed Mahameed
---
drivers/net/ethernet/mellanox/mlx5/core/en_rep.
From: Gavi Teitz
Unify and isolate the error handling flow in mlx5_mpfs_add_mac(),
removing code duplication.
Signed-off-by: Gavi Teitz
Signed-off-by: Saeed Mahameed
---
.../ethernet/mellanox/mlx5/core/lib/mpfs.c| 26 +++
1 file changed, 15 insertions(+), 11 deletions(-)
From: Ariel Levkovich
Set supported device features in the netdevice MPLS features mask.
This will enable HW checksumming and TSO for MPLS tagged traffic.
Signed-off-by: Ariel Levkovich
Signed-off-by: Saeed Mahameed
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 5 +
1 file chang
From: Gavi Teitz
Remove the limitation preventing adding a vport's MAC address to the
Multi-Physical Function Switch (MPFS) more than once per E-switch, as
there is no difference in the MPFS if an address is being used by an
E-switch more than once.
This allows the E-switch to have multiple vpor
From: Parav Pandit
Consider PCI and non PCI device types while setting device name
in get_drvinfo() callback using existing generic device.
Signed-off-by: Parav Pandit
Reviewed-by: Vu Pham
Signed-off-by: Saeed Mahameed
---
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 2 +-
1 file c
From: Ariel Levkovich
This patch changes the way the driver advertises its checksum offload
capabilities within the net device features bit mask.
Instead of advertising protocol specific checksumming capabilities
which are limited today to IPv4 and IPv6, we move to reporing
generic HW checksummi
Commit 86029d10af18 ("tls: zero the crypto information from tls_context
before freeing") added memzero_explicit() calls to clear the key material
before freeing struct tls_context, but it missed tls_device.c has its
own way of freeing this structure. Replace the missing free.
Fixes: 86029d10af18 (
Make sure that wide stores are allowed at proper (aligned) addresses.
Note that user_ip6 is naturally aligned on 8-byte boundary, so
correct addresses are user_ip6[0] and user_ip6[2]. msg_src_ip6 is,
however, aligned on a 4-byte bondary, so only msg_src_ip6[1]
can be wide-stored.
Cc: Andrii Nakryi
Since commit cd17d7770578 ("bpf/tools: sync bpf.h") clang decided
that it can do a single u64 store into user_ip6[2] instead of two
separate u32 ones:
# 17: (18) r2 = 0x100
# ; ctx->user_ip6[2] = bpf_htonl(DST_REWRITE_IP6_2);
# 19: (7b) *(u64 *)(r1 +16) = r2
# invalid bpf_conte
Neither drivers nor the tls offload code currently supports TLS
version 1.3. Check the TLS version when installing connection
state. TLS 1.3 will just fallback to the kernel crypto for now.
Fixes: 130b392c6cd6 ("net: tls: Add tls 1.3 support")
Signed-off-by: Jakub Kicinski
Reviewed-by: Dirk van d
This patch series adds the gve driver which will support the
Compute Engine Virtual NIC that will be available in the future.
v2:
- Patch 1:
- Remove gve_size_assert.h and use static_assert instead.
- Loop forever instead of bugging if the device won't reset
- Use module_pci_driver
- Patch 2
Add support for the following ethtool commands:
ethtool -s|--change devname [msglvl N] [msglevel type on|off]
ethtool -S|--statistics devname
ethtool -i|--driver devname
ethtool -l|--show-channels devname
ethtool -L|--set-channels devname
ethtool -g|--show-ring devname
ethtool --reset devname
Sig
On 6/28/19 10:03 AM, Andrea Claudi wrote:
> As parse_percent() is used only in tc.
>
> This reduces ip, bridge and genl binaries size:
>
> $ bloat-o-meter -t bridge/bridge bridge/bridge.new
> add/remove: 0/1 grow/shrink: 0/0 up/down: 0/-109 (-109)
> Total: Before=50973, After=50864, chg -0.21%
>
Add a driver framework for the Compute Engine Virtual NIC that will be
available in the future.
At this point the only functionality is loading the driver.
Signed-off-by: Catherine Sullivan
Signed-off-by: Sagi Shahar
Signed-off-by: Jon Olson
Acked-by: Willem de Bruijn
Reviewed-by: Luigi Rizzo
Add support for passing traffic.
Signed-off-by: Catherine Sullivan
Signed-off-by: Sagi Shahar
Signed-off-by: Jon Olson
Acked-by: Willem de Bruijn
Reviewed-by: Luigi Rizzo
---
.../networking/device_drivers/google/gve.rst | 30 +
drivers/net/ethernet/google/gve/Makefile | 2 +-
driver
Add support for the workqueue to handle management interrupts and
support for resets.
Signed-off-by: Catherine Sullivan
Signed-off-by: Sagi Shahar
Signed-off-by: Jon Olson
Acked-by: Willem de Bruijn
Reviewed-by: Luigi Rizzo
---
.../networking/device_drivers/google/gve.rst | 11 +
drivers/n
On Fri, Jun 28, 2019 at 1:06 PM Andrew Lunn wrote:
>
> On Fri, Jun 28, 2019 at 11:46:15AM -0700, Jakub Kicinski wrote:
> > On Fri, 28 Jun 2019 10:52:27 -0700, Catherine Sullivan wrote:
> > > > > +if NET_VENDOR_GOOGLE
> > > > > +
> > > > > +config GVE
> > > > > + tristate "Google Virtual NIC (g
On Fri, Jun 28, 2019 at 11:11 AM Stephen Hemminger
wrote:
>
> On Fri, 28 Jun 2019 10:56:33 -0700
> Catherine Sullivan wrote:
>
> > +static void
> > +gve_get_ethtool_stats(struct net_device *netdev,
> > + struct ethtool_stats *stats, u64 *data)
> > +{
> > + struct gve_priv *p
On 6/24/19 10:34 PM, Hoang Le wrote:
> Support for indicating interface name has an ip address in parallel
> with specifying ip address when activating UDP bearer.
> This liberates the user from keeping track of the current ip address
> for each device.
>
> Old command syntax:
> $tipc bearer enabl
On 6/28/19 3:49 PM, Jeff Kirsher wrote:
From: Dann Frazier
An ipsec structure will not be allocated if the hardware does not support
offload. Fixes the following Oops:
[ 191.045452] Unable to handle kernel NULL pointer dereference at virtual
address
[ 191.054232] Mem abort
From: Colin Ian King
A recent commit efa14c3985828d ("iavf: allow null RX descriptors") added
a null pointer sanity check on rx_buffer, however, rx_buffer is being
dereferenced before that check, which implies a null pointer dereference
bug can potentially occur. Fix this by only dereferencing r
From: Artem Bityutskiy
This patch adds the RR2DCDELAY register to the ethtool registers dump.
RR2DCDELAY exists on I210 and I211 Intel Gigabit Ethernet chips and it stands
for "Read Request To Data Completion Delay". Here is how this register is
described in the I210 datasheet:
"This field captu
From: Vitaly Lifshits
Due to commit: 5d8682588605 ("[misc] mei: me: allow runtime
pm for platform with D0i3")
When disconnecting the cable and reconnecting it the NIC
enters DMoff state. This caused wrong link indication
and duplex mismatch. This bug is described in:
https://bugzilla.redhat.com/s
On Fri, 28 Jun 2019 12:40:29 -0700, John Fastabend wrote:
> The lock() is already held when entering unhash() side so need to
> handle this case as well,
>
> CPU 0 (free) CPU 1 (wq)
>
> lock(sk) ctx = tls_get_ctx(sk) <- need to be check null ptr
> sk_prot->unhash()
> set_b
This aligns the iavf_debug() macro with the other Intel drivers.
Add the bus number, bus_id field to i40e_bus_info so output shows
each physical port(i.e func) in following format:
]:]]:][][.[]]
domains are numbered from 0 to ), bus (0-ff), slot (0-1f) and
function (0-7).
Signed-off-by:
From: "Gustavo A. R. Silva"
Make use of the struct_size() helper instead of an open-coded version
in order to avoid any potential type mistakes, in particular in the
context in which this code is being used.
So, replace code of the following form:
sizeof(struct virtchnl_ether_addr_list) + (coun
From: Colin Ian King
The u32 variable rem is being shifted using u32 arithmetic however
it is being passed to div_u64 that expects the expression to be a u64.
The 32 bit shift may potentially overflow, so cast rem to a u64 before
shifting to avoid this. Also remove comment about overflow.
Addre
This series contains a smorgasbord of updates to many of the Intel
drivers.
Gustavo A. R. Silva updates the ice and iavf drivers to use the
strcut_size() helper where possible.
Miguel increases the pause and refresh time for flow control in the
e1000e driver during reset for certain devices.
Dan
From: Miguel Bernal Marin
Suggested-by: Tim Pepper
Signed-off-by: Miguel Bernal Marin
Signed-off-by: Paul Menzel
Acked-by: Sasha Neftin
Tested-by: Aaron Brown
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/e1000e/netdev.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-
From: Detlev Casanova
Use delayed work instead of timers to run the watchdog of the e1000e
driver.
Simplify the code with one less middle function.
Signed-off-by: Detlev Casanova
Tested-by: Aaron Brown
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/e1000e/e1000.h | 5 +-
drive
From: Dann Frazier
An ipsec structure will not be allocated if the hardware does not support
offload. Fixes the following Oops:
[ 191.045452] Unable to handle kernel NULL pointer dereference at virtual
address
[ 191.054232] Mem abort info:
[ 191.057014] ESR = 0x9604
[
From: Arjan van de Ven
The e1000e driver is a great user of the usleep_range() API,
and has nice ranges that in principle help power management.
However the ranges that are used only during system startup are
very long (and can add easily 100 msec to the boot time) while
the power savings of suc
From: Artem Bityutskiy
This patch has no functional impact and it is just a preparation
for the following patch. It removes an early return from the
'igb_get_regs()' function by moving the 82576-only registers
dump into an "if" block. With this preparation, we can dump more
non-82576 registers at
From: Alexander Duyck
Change the ethtool link settings call to just read the cached state out of
the adapter structure instead of trying to recheck the value from the PF.
Doing this should prevent excessive reading of the mailbox.
Signed-off-by: Alexander Duyck
Reviewed-by: "Guilherme G. Piccol
From: Venkatesh Srinivas
e1000 writes to doorbells to post transmit descriptors and fill the
receive ring. After writing descriptors to memory but before
writing to doorbells, use dma_wmb() rather than wmb(). wmb() is more
heavyweight than necessary for a device to see descriptor writes.
On x86,
From: Harshitha Ramamurthy
This patch enables macvlan offloads for i40e. The idea is to use
channels as macvlan interfaces. The channels are VSIs of
type VMDQ. When the first macvlan is created, the maximum number of
channels possible are created. From then on, as a macvlan interface
is created,
From: "Gustavo A. R. Silva"
One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:
struct foo {
int stuff;
struct boo entry[];
};
s
On Fri, Jun 28, 2019 at 11:22 AM Michael Chan wrote:
>
> On Thu, Jun 27, 2019 at 12:42 PM Mahesh Bandewar wrote:
>
> > However, Michael Chan had a setup
> > where these fixes helped him mitigate the issue and not cause
> > the crash.
> >
>
> Our lab has finished testing these patches. The patch
From: Huy Nguyen
Rename mlx5_pci_dev_type to mlx5_coredev_type to distinguish different mlx5
device types.
mlx5_coredev_type represents mlx5_core_dev instance type. Hence keep
mlx5_coredev_type in mlx5_core_dev structure.
Signed-off-by: Huy Nguyen
Signed-off-by: Vu Pham
Signed-off-by: Parav P
From: Bodong Wang
Use the defined iterators to traversal VF reps/vport. Also, rely on
num of VFs rather than the counter of enabled vports as PF will also
be enabled from ECPF side, and the counter will be different from
num of VFs.
Signed-off-by: Bodong Wang
Reviewed-by: Parav Pandit
Signed-o
From: Bodong Wang
When NVME device emulation mode is enabled, more than one PFs use the
same physical port. In this case, MPFS is required to program L2
addresses.
It used to rely on netdev set_rx_mode in switchdev mode, but driver
later changed to not create netdev for eswitch manager once in
s
From: Bodong Wang
When ECPF is the eswitch manager, host PF is treated like other VFs.
Driver should do the same for inline mode and vlan pop.
Add new iterators to include host PF if ECPF is the eswitch manager.
Signed-off-by: Bodong Wang
Signed-off-by: Saeed Mahameed
---
.../net/ethernet/me
From: Bodong Wang
When driver is doing eswitch mode change, it's critical to keep number
of enabled VFs unchanged. However, it can be changed on the fly once
function changed event is registered.
To remove this uncertainty, function changed event should not be
registered before all setups, and f
From: Parav Pandit
As mlx5_get_next_phys_dev is used only for PCI PF devices use case,
limit it to search only for PCI devices.
Signed-off-by: Parav Pandit
Reviewed-by: Vu Pham
Signed-off-by: Saeed Mahameed
---
drivers/net/ethernet/mellanox/mlx5/core/dev.c | 9 -
1 file changed, 8 in
From: Bodong Wang
Driver is referring to the array index when doing rep initialization,
using vport is confusing as it's normally interpreted as vport number.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang
Reviewed-by: Parav Pandit
Reviewed-by: Mark Bloch
Signed-off-
From: Shay Agroskin
Given a fw component index, the MCQI register allows us to query
this component's information (e.g. its version and capabilities).
Given a fw component index, the MCQS register allows us to query the
status of a fw component, including its type and state
(e.g. PRESET/IN_USE).
From: Bodong Wang
When an IB rep is loaded, netdev for the same vport is saved for later
reference. However, it's not cleaned up when doing unload. For ECPF,
kernel crashes when driver is referring to the already removed netdev.
Following steps lead to a shown call trace:
1. Create n VFs from ho
From: Bodong Wang
Devlink eswitch mode is not necessarily related to SR-IOV, e.g, ECPF
can be at offload mode when SR-IOV is not enabled.
Rename the interface and eswitch mode names to decouple from SR-IOV,
and cleanup eswitch messages accordingly.
Signed-off-by: Bodong Wang
Signed-off-by: Sae
From: Bodong Wang
Enabled number of VFs is key for eswich manager to do flow steering
initialization and vport configurations. However, the number of
enabled VFs may come from two sources as below.
PF: num of VFs is provided by enabled SR-IOV of itself.
ECPF: num of VFs is provided by enabled SR
From: Parav Pandit
mlx5_pci_init() performs pci specific initialization of the
mlx5_core_dev struct.
Hence move pci_status_mutex to pci initialization routine
mlx5_pci_init().
This allows reusing mlx5_mdev_init() to non PCI devices.
Signed-off-by: Parav Pandit
Reviewed-by: Vu Pham
Signed-off-b
From: Parav Pandit
Update mlx5 device interface data structures for:
1. New command definitions for allocating, deallocating SF
2. Query SF partition
3. Eswitch SF fields
4. HCA CAP SF fields
5. Extend Eswitch functions command for SF
Signed-off-by: Parav Pandit
Signed-off-by: Vu Pham
Signed-o
From: Bodong Wang
Before the offending commit, vlan will be configured if either vlan
or qos is set. After the change with new set flags, function callers
should provide flags accordingly.
Fixes: e33dfe316cf3 ("net/mlx5: E-Switch, Allow fine tuning of eswitch vport
push/pop vlan")
Signed-off-by
From: Bodong Wang
When ECPF is eswitch manager, it has the privilege to query and
configure the mac and node guid of host PF.
While vport number of host PF is 0, the vport command should be
issued with other_vport set in this case as the cmd is issued by
ECPF vport(0xfffe).
Add a specific funct
Hi All,
This series includes some low level updates mainly in the E-Switch
netdev and rdma vport representors areas.
From Parav and Huy:
1) Added hardware bits and structures definitions for sub-functions
2) Small code cleanup and improvement for PF pci driver.
From Bodong:
3) Use the correct
From: Bodong Wang
When ECPF eswitch manager is at offloads mode, it monitors functions
changed event from host PF side and acts according to the number of
VFs enabled/disabled.
As ECPF and host PF work in two independent hosts, it's possible that
host PF OS reboots but ECPF system is still kept
From: Parav Pandit
While enabling SR-IOV, PCI core already checks that if SR-IOV is already
enabled, it returns failure error code.
Hence, remove such duplicate check from mlx5_core driver.
While at it, make mlx5_device_disable_sriov() to perform cleanup of VFs in
reverse order of mlx5_device_en
From: Bodong Wang
In the single IB device mode, the mapping between vport number and
rep relies on a counter. However for dynamic vport allocation, it is
desired to keep consistent map of eswitch vport and IB port.
Hence, simplify code to remove the free running counter and instead
use the avail
On Thu, 27 Jun 2019 21:37:18 +0300
Baruch Siach wrote:
> 32bit targets define uint64_t as long long unsigned. This leads to the
> following build warning:
>
> devlink.c: In function ‘pr_out_u64’:
> devlink.c:1729:11: warning: format ‘%lu’ expects argument of type ‘long
> unsigned int’, but argu
On Tue, 25 Jun 2019 12:29:54 +0200
Andrea Claudi wrote:
> 'home', 'nodad' and 'mngtmpaddr' options are IPv6-only, but
> it is possible to set them on IPv4 addresses, too. This should
> not be possible.
>
> Fix this adding a check on the protocol family before setting
> the flags, and print warni
When the XDP program attached to a zero-copy AF_XDP socket returns XDP_TX,
queue the umem frame on the XDP TX ring. Space on the recycle stack is
pre-allocated when the xsk is created. (taken from tx_ring, since the
xdp ring is not initialized yet)
Signed-off-by: Jonathan Lemon
---
drivers/ne
When the XDP program attached to a zero-copy AF_XDP socket returns XDP_TX,
queue the umem frame on the XDP TX ring. Space on the recycle stack is
pre-allocated when the xsk is created. (taken from tx_ring, since the
xdp ring is not initialized yet)
Signed-off-by: Jonathan Lemon
---
drivers/net
1 - 100 of 300 matches
Mail list logo
