* Re: [PATCH net-next 4/6] net: dsa: mv88e6xxx: do not change STP state on port disabling
From: Vivien Didelot @ 2019-08-19 15:27 UTC (permalink / raw)
To: Andrew Lunn; +Cc: netdev, marek.behun, davem, f.fainelli
In-Reply-To: <20190819134057.GF8981@lunn.ch>
Hi Andrew,
On Mon, 19 Aug 2019 15:40:57 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> On Sun, Aug 18, 2019 at 01:35:46PM -0400, Vivien Didelot wrote:
> > When disabling a port, that is not for the driver to decide what to
> > do with the STP state. This is already handled by the DSA layer.
>
> Hi Vivien
>
> Putting the port into STP disabled state is how you actually disable
> it, for the mv88e6xxx. So this is not really about STP, it is about
> powering off the port. Maybe a comment is needed, rather than removing
> the code?
This is not for the driver to decide, the stack already handles that.
Otherwise, calling dsa_port_disable on a bridged port would result in
mv88e6xxx forcing the STP state to Disabled while this is not expected.
Thanks,
Vivien
^ permalink raw reply
* Re: [PATCH 1/2] PTP: introduce new versions of IOCTLs
From: Richard Cochran @ 2019-08-19 15:43 UTC (permalink / raw)
To: Joe Perches; +Cc: Felipe Balbi, Christopher S Hall, netdev, linux-kernel
In-Reply-To: <83075553a61ede1de9cbf77b90a5acdeab5aacbf.camel@perches.com>
On Sun, Aug 18, 2019 at 03:07:18PM -0700, Joe Perches wrote:
> Also the original patch deletes 2 case entries for
> PTP_PIN_GETFUNC and PTP_PIN_SETFUNC and converts them to
> PTP_PIN_GETFUNC2 and PTP_PIN_SETFUNC2 but still uses tests
> for the deleted case label entries making part of the case
> code block unreachable.
>
> That's at least a defect:
>
> - case PTP_PIN_GETFUNC:
> + case PTP_PIN_GETFUNC2:
>
> and
>
> - case PTP_PIN_SETFUNC:
> + case PTP_PIN_SETFUNC2:
Good catch. Felipe, please fix that!
(Regarding Joe's memset suggestion, I'll leave that to your discretion.)
Thanks,
Richard
^ permalink raw reply
* Re: [RESEND][PATCH v3 bpf-next] btf: expose BTF info through sysfs
From: Ilya Leoshkevich @ 2019-08-19 15:56 UTC (permalink / raw)
To: Andrii Nakryiko
Cc: bpf, Networking, Alexei Starovoitov, Daniel Borkmann,
Andrii Nakryiko, Kernel Team, Vasily Gorbik, Heiko Carstens
In-Reply-To: <20190812183947.130889-1-andriin@fb.com>
> Am 12.08.2019 um 20:39 schrieb Andrii Nakryiko <andriin@fb.com>:
>
> @@ -92,23 +92,34 @@ vmlinux_link()
> }
>
> # generate .BTF typeinfo from DWARF debuginfo
> +# ${1} - vmlinux image
> +# ${2} - file to dump raw BTF data into
> gen_btf()
> {
> - local pahole_ver;
> + local pahole_ver
> + local bin_arch
>
> if ! [ -x "$(command -v ${PAHOLE})" ]; then
> info "BTF" "${1}: pahole (${PAHOLE}) is not available"
> - return 0
> + return 1
> fi
>
> pahole_ver=$(${PAHOLE} --version | sed -E 's/v([0-9]+)\.([0-9]+)/\1\2/')
> if [ "${pahole_ver}" -lt "113" ]; then
> info "BTF" "${1}: pahole version $(${PAHOLE} --version) is too old, need at least v1.13"
> - return 0
> + return 1
> fi
>
> - info "BTF" ${1}
> + info "BTF" ${2}
> + vmlinux_link ${1}
> LLVM_OBJCOPY=${OBJCOPY} ${PAHOLE} -J ${1}
> +
> + # dump .BTF section into raw binary file to link with final vmlinux
> + bin_arch=$(${OBJDUMP} -f ${1} | grep architecture | \
> + cut -d, -f1 | cut -d' ' -f2)
> + ${OBJCOPY} --dump-section .BTF=.btf.kernel.bin ${1} 2>/dev/null
> + ${OBJCOPY} -I binary -O ${CONFIG_OUTPUT_FORMAT} -B ${bin_arch} \
> + --rename-section .data=.BTF .btf.kernel.bin ${2}
> }
CONFIG_OUTPUT_FORMAT appears to be x86-only; enabling
CONFIG_DEBUG_INFO_BTF on s390 caused a build failure. I now have a quick
and dirty local patch, which adds CONFIG_OUTPUT_FORMAT to s390 and fixes
the issue for me, but I suspect that CONFIG_DEBUG_INFO_BTF might be
broken on other arches as well.
^ permalink raw reply
* Re: [PATCH net-next] netdevsim: Fix build error without CONFIG_INET
From: Ido Schimmel @ 2019-08-19 15:58 UTC (permalink / raw)
To: YueHaibing
Cc: davem, idosch, jiri, mcroce, linux-kernel, netdev, jakub.kicinski
In-Reply-To: <20190819120825.74460-1-yuehaibing@huawei.com>
On Mon, Aug 19, 2019 at 08:08:25PM +0800, YueHaibing wrote:
> If CONFIG_INET is not set, building fails:
>
> drivers/net/netdevsim/dev.o: In function `nsim_dev_trap_report_work':
> dev.c:(.text+0x67b): undefined reference to `ip_send_check'
>
> Add CONFIG_INET Kconfig dependency to fix this.
>
> Reported-by: Hulk Robot <hulkci@huawei.com>
> Fixes: da58f90f11f5 ("netdevsim: Add devlink-trap support")
> Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Thanks for the patch.
> ---
> drivers/net/Kconfig | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
> index 48e209e..7bb786e 100644
> --- a/drivers/net/Kconfig
> +++ b/drivers/net/Kconfig
> @@ -505,7 +505,7 @@ source "drivers/net/hyperv/Kconfig"
>
> config NETDEVSIM
> tristate "Simulated networking device"
> - depends on DEBUG_FS
> + depends on INET && DEBUG_FS
> select NET_DEVLINK
> help
> This driver is a developer testing tool and software model that can
> --
> 2.7.4
>
>
^ permalink raw reply
* Re: [PATCH 1/2] PTP: introduce new versions of IOCTLs
From: Joe Perches @ 2019-08-19 15:58 UTC (permalink / raw)
To: Richard Cochran; +Cc: Felipe Balbi, Christopher S Hall, netdev, linux-kernel
In-Reply-To: <20190819154320.GB2883@localhost>
On Mon, 2019-08-19 at 08:43 -0700, Richard Cochran wrote:
> On Sun, Aug 18, 2019 at 03:07:18PM -0700, Joe Perches wrote:
> > Also the original patch deletes 2 case entries for
> > PTP_PIN_GETFUNC and PTP_PIN_SETFUNC and converts them to
> > PTP_PIN_GETFUNC2 and PTP_PIN_SETFUNC2 but still uses tests
> > for the deleted case label entries making part of the case
> > code block unreachable.
> >
> > That's at least a defect:
> >
> > - case PTP_PIN_GETFUNC:
> > + case PTP_PIN_GETFUNC2:
> >
> > and
> >
> > - case PTP_PIN_SETFUNC:
> > + case PTP_PIN_SETFUNC2:
>
> Good catch. Felipe, please fix that!
>
> (Regarding Joe's memset suggestion, I'll leave that to your discretion.)
Not just how declarations are done or memset.
Minimizing unnecessary stack consumption is generally good.
^ permalink raw reply
* Re: [PATCH v4 3/9] nvmem: core: add nvmem_device_find
From: Srinivas Kandagatla @ 2019-08-19 16:03 UTC (permalink / raw)
To: Thomas Bogendoerfer
Cc: Ralf Baechle, Paul Burton, James Hogan, Dmitry Torokhov,
Lee Jones, David S. Miller, Alessandro Zummo, Alexandre Belloni,
Greg Kroah-Hartman, Jiri Slaby, Evgeniy Polyakov, linux-mips,
linux-kernel, linux-input, netdev, linux-rtc, linux-serial
In-Reply-To: <20190816140942.GA15050@alpha.franken.de>
On 16/08/2019 15:09, Thomas Bogendoerfer wrote:
> On Wed, Aug 14, 2019 at 01:52:49PM +0100, Srinivas Kandagatla wrote:
>> On 14/08/2019 12:46, Thomas Bogendoerfer wrote:
>>> On Tue, 13 Aug 2019 10:40:34 +0100
>>> Srinivas Kandagatla <srinivas.kandagatla@linaro.org> wrote:
>>>> On 09/08/2019 11:32, Thomas Bogendoerfer wrote:
>>>>> nvmem_device_find provides a way to search for nvmem devices with
>>>>> the help of a match function simlair to bus_find_device.
>>>>>
>>>>> Signed-off-by: Thomas Bogendoerfer <tbogendoerfer@suse.de>
>>>>> ---
>>>>> drivers/nvmem/core.c | 62 ++++++++++++++++++++++--------------------
>>>>> include/linux/nvmem-consumer.h | 9 ++++++
>>>>> 2 files changed, 41 insertions(+), 30 deletions(-)
>>>>
>>>> Have you considered using nvmem_register_notifier() ?
>>>
>>> yes, that was the first idea. But then I realized I need to build up
>>> a private database of information already present in nvmem bus. So I
>>> looked for a way to retrieve it from there. Unfortunately I couldn't
>>> use bus_find_device directly, because nvmem_bus_type and struct nvmem_device
>>> is hidden. So I refactured the lookup code and added a more universal
>>> lookup function, which fits my needs and should be usable for more.
>> I see your point.
>>
>> overall the patch as it is look good, but recently we added more generic
>> lookups for DT node, looks like part of your patch is un-doing generic
>> device name lookup.
>>
>> DT node match lookup is in https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git/log/?h=generic_lookup_helpers
>
> these patches are not in Linus tree, yet. I guess they will show up
> in 5.4. No idea how to deal with it right now, do you ?
All these patches are due to go in next merge window,
You should base your patch on top of linux-next.
thanks,
srini
>
>> Other missing bit is adding this api to documentation in
>> ./Documentation/driver-api/nvmem.rst
>
> ok, will do.
>
> Thomas.
>
^ permalink raw reply
* Re: r8169: Performance regression and latency instability
From: Juliana Rodrigueiro @ 2019-08-19 16:04 UTC (permalink / raw)
To: Heiner Kallweit, Holger Hoffstätte, Eric Dumazet, netdev
In-Reply-To: <8fa71d82-d309-df38-5924-2275db188b61@gmail.com>
Hi!
First of all: Thank you everyone for the input.
Here is some more info about my NIC. (Using the latest ethtool)
# ethtool -i eth0 ; ifconfig eth0
driver: r8169
version:
firmware-version: rtl8168h-2_0.0.2 02/26/15
expansion-rom-version:
bus-info: 0000:02:00.0
supports-statistics: yes
supports-test: no
supports-eeprom-access: no
supports-register-dump: yes
supports-priv-flags: no
eth0 Link encap:Ethernet HWaddr <hidden>
inet addr:<hidden> Bcast:<hidden> Mask:255.255.0.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:27392501 errors:0 dropped:0 overruns:0 frame:0
TX packets:24647212 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:33702173568 (31.3 GiB) TX bytes:35865124147 (33.4 GiB)
On 8/16/19 9:12 PM, Heiner Kallweit wrote:
> Indeed, here we're talking about changes in linux-next, and Juliana's issue is
> with 4.19. However I'd appreciate if Juliana could test with linux-next and
> different combinations of the NETIF_F_xxx features.
I also tested the latest linux-next (20190819) and the results did not
improved for me, unfortunately. About the same as all the kernel
versions I tested from 4.17 onwards.
# netperf -v 2 -P 0 -H <netserver-ip>,4 -I 99,5 -t omni -l 1 -- -O
STDDEV_LATENCY -m 64K -d Send
627.99
Running linux-next I have the following defaults (shortened for simplicity):
# ethtool -k eth0
Features for eth0:
rx-checksumming: on
tx-checksumming: on
tx-checksum-ipv4: on
tx-checksum-ip-generic: off [fixed]
tx-checksum-ipv6: on
tx-checksum-fcoe-crc: off [fixed]
tx-checksum-sctp: off [fixed]
scatter-gather: on
tx-scatter-gather: on
tx-scatter-gather-fraglist: off [fixed]
tcp-segmentation-offload: on
tx-tcp-segmentation: on
tx-tcp-ecn-segmentation: off [fixed]
tx-tcp-mangleid-segmentation: off
tx-tcp6-segmentation: on
generic-segmentation-offload: on
generic-receive-offload: on
large-receive-offload: off [fixed]
rx-vlan-offload: on
tx-vlan-offload: on
... (all off from here)
There are quite a few possible combinations to go through. I executed my
test with SG, TSO, GSO, RX, TX individually disabled, but the results
were all the same or slightly worse.
Until I find the root cause, we will have to keep the "tcp: switch to
GSO being always on" patch reverted for production, which is not ideal.
Any other ideas how I could debug this issue?
Best regards,
Juliana.
^ permalink raw reply
* Re: [PATCH net-next 4/6] net: dsa: mv88e6xxx: do not change STP state on port disabling
From: Andrew Lunn @ 2019-08-19 16:10 UTC (permalink / raw)
To: Vivien Didelot; +Cc: netdev, marek.behun, davem, f.fainelli
In-Reply-To: <20190819112733.GD6123@t480s.localdomain>
On Mon, Aug 19, 2019 at 11:27:33AM -0400, Vivien Didelot wrote:
> Hi Andrew,
>
> On Mon, 19 Aug 2019 15:40:57 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> > On Sun, Aug 18, 2019 at 01:35:46PM -0400, Vivien Didelot wrote:
> > > When disabling a port, that is not for the driver to decide what to
> > > do with the STP state. This is already handled by the DSA layer.
> >
> > Hi Vivien
> >
> > Putting the port into STP disabled state is how you actually disable
> > it, for the mv88e6xxx. So this is not really about STP, it is about
> > powering off the port. Maybe a comment is needed, rather than removing
> > the code?
>
> This is not for the driver to decide, the stack already handles that.
> Otherwise, calling dsa_port_disable on a bridged port would result in
> mv88e6xxx forcing the STP state to Disabled while this is not expected.
Hi Vivien
Lets look at this from a different angle.
The chip powers up. The older generation, the ports are enabled by
default. For newer generations the NO_CPU strapping determines the
power up state of a port.
What we want is that unused ports get powered off. The previous change
in the set does this. It calls mv88e6xxx_port_disable for all unused
ports. How do we disable the port? For the mv88ex666, we set the port
state to disabled. "The switch port is disabled and it will not
receive or transmit and frames." For other switches, there might be
other control registers to play with, other than STP.
Are you saying the core already sets the STP to disabled, for ports
which are unused? I did not spot that in your previous patch?
Thanks
Andrew
^ permalink raw reply
* [PATCH] bpfilter/verifier: add include guard to tnum.h
From: Masahiro Yamada @ 2019-08-19 16:10 UTC (permalink / raw)
To: Alexei Starovoitov, Daniel Borkmann, Martin KaFai Lau, Song Liu,
Yonghong Song, netdev, bpf
Cc: Masahiro Yamada, linux-kernel
Add a header include guard just in case.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---
include/linux/tnum.h | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/include/linux/tnum.h b/include/linux/tnum.h
index c7dc2b5902c0..c17af77f3fae 100644
--- a/include/linux/tnum.h
+++ b/include/linux/tnum.h
@@ -5,6 +5,10 @@
* propagate the unknown bits such that the tnum result represents all the
* possible results for possible values of the operands.
*/
+
+#ifndef _LINUX_TNUM_H
+#define _LINUX_TNUM_H
+
#include <linux/types.h>
struct tnum {
@@ -81,3 +85,5 @@ bool tnum_in(struct tnum a, struct tnum b);
int tnum_strn(char *str, size_t size, struct tnum a);
/* Format a tnum as tristate binary expansion */
int tnum_sbin(char *str, size_t size, struct tnum a);
+
+#endif /* _LINUX_TNUM_H */
--
2.17.1
^ permalink raw reply related
* [net-next v2 00/14][pull request] 100GbE Intel Wired LAN Driver Updates 2019-08-19
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, nhorman, sassmann
This series contains updates to ice driver only.
Brett fixes the detection of a hung transmit ring by checking the
software based tail (next_to_use) to determine if there is pending work.
Updates the driver to assume that using more than one receive queue per
receive ring container is a rare case, so use unlikely() in the case
were we actually need to divide our budget for multiple queues. Fixed
an issue where the write back on ITR bit was not being set when
interrupts are disabled, which was causing only write backs when polling
only when a cache line is filled. Cleans up unnecessary wait times
during VF bring up and reset paths. Increased the mailbox size for
receive queues that are used to communicate with VFs to accommodate the
large number of VFs that the driver can support.
Akeem restructures the initialization flows for VFs, including how VFs
are configured and resources allocated to improve flows so that when we
clean up resources, we do not try to free resources that were never
allocated. Organizes code to ensure that VF specific code is located in
the SR-IOV specific file.
Paul fixes an issue when setting the pause parameter which was
incorrectly blocking users from changing receive or transmit pause
settings. Ensure register access for MSIX vector index is only done in
the PF space and not absolute device space.
Usha fixes a potential kernel hang in the DCB rebuild path when in CEE
mode, where the ETS recommended DCB configuration is not being set or
set correctly.
Mitch updates the driver to process all receive descriptors, regardless
of the size of the associated data.
Tony fixes and issue during the reset/rebuild path of a PF VSI where we
were assuming that the PF VSI was always to be enabled, which can
attempt to bring up a PF VSI on a downed interface which can lead to
various crashes.
Pawel fixes up variable definitions to match the type of data being
stored.
v2: Dropped patch 1 of the series to add ethtool support to query/add
channels on a VSI, while we re-qork the functionality to match the
ethtool expected behavior to report combined (Tx and Rx) numbers.
The following are changes since commit 20e79a0a2cfd15b6cfb18119f2e108396be56716:
net: hns: add phy_attached_info() to the hns driver
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue 100GbE
Akeem G Abodunrin (2):
ice: Restructure VFs initialization flows
ice: Move VF resources definition to SR-IOV specific file
Brett Creeley (6):
ice: Use the software based tail when checking for hung Tx ring
ice: Assume that more than one Rx queue is rare in ice_napi_poll
ice: Set WB_ON_ITR when we don't re-enable interrupts
ice: Reduce wait times during VF bringup/reset
ice: Increase size of Mailbox receive queue for many VFs
ice: improve print for VF's when adding/deleting MAC filters
Mitch Williams (1):
ice: allow empty Rx descriptors
Paul Greenwalt (2):
ice: fix set pause param autoneg check
ice: update GLINT_DYN_CTL and GLINT_VECT2FUNC register access
Pawel Kaminski (1):
ice: Change type for queue counts
Tony Nguyen (1):
ice: Do not always bring up PF VSI in ice_ena_vsi()
Usha Ketineni (1):
ice: Fix kernel hang with DCB reset in CEE mode
drivers/net/ethernet/intel/ice/ice.h | 14 +-
drivers/net/ethernet/intel/ice/ice_dcb_lib.c | 149 +++++++++++-------
drivers/net/ethernet/intel/ice/ice_ethtool.c | 28 +++-
.../net/ethernet/intel/ice/ice_hw_autogen.h | 3 +
drivers/net/ethernet/intel/ice/ice_main.c | 12 +-
drivers/net/ethernet/intel/ice/ice_txrx.c | 84 +++++++++-
drivers/net/ethernet/intel/ice/ice_txrx.h | 13 ++
.../net/ethernet/intel/ice/ice_virtchnl_pf.c | 148 ++++++++++-------
.../net/ethernet/intel/ice/ice_virtchnl_pf.h | 20 ++-
9 files changed, 325 insertions(+), 146 deletions(-)
--
2.21.0
^ permalink raw reply
* [net-next v2 02/14] ice: Assume that more than one Rx queue is rare in ice_napi_poll
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem; +Cc: Brett Creeley, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
Currently we divide budget by the number of Rx queues per Rx ring
container in ice_napi_poll even if there is only 1. This is an
unnecessary divide for the normal case of 1 Rx ring per Rx ring
container. Fix this by using an unlikely() call in the case where we
actually need to divide.
Also, we will always set budget_per_ring even if there are no Rx rings
in the Rx ring container so we don't need to initialize it to 0.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_txrx.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index 9234fd203929..837d6ae2f33b 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -1414,8 +1414,8 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
container_of(napi, struct ice_q_vector, napi);
struct ice_vsi *vsi = q_vector->vsi;
bool clean_complete = true;
- int budget_per_ring = 0;
struct ice_ring *ring;
+ int budget_per_ring;
int work_done = 0;
/* Since the actual Tx work is minimal, we can give the Tx a larger
@@ -1429,11 +1429,16 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
if (budget <= 0)
return budget;
- /* We attempt to distribute budget to each Rx queue fairly, but don't
- * allow the budget to go below 1 because that would exit polling early.
- */
- if (q_vector->num_ring_rx)
+ /* normally we have 1 Rx ring per q_vector */
+ if (unlikely(q_vector->num_ring_rx > 1))
+ /* We attempt to distribute budget to each Rx queue fairly, but
+ * don't allow the budget to go below 1 because that would exit
+ * polling early.
+ */
budget_per_ring = max(budget / q_vector->num_ring_rx, 1);
+ else
+ /* Max of 1 Rx ring in this q_vector so give it the budget */
+ budget_per_ring = budget;
ice_for_each_ring(ring, q_vector->rx) {
int cleaned;
--
2.21.0
^ permalink raw reply related
* [net-next v2 07/14] ice: allow empty Rx descriptors
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem
Cc: Mitch Williams, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Mitch Williams <mitch.a.williams@intel.com>
In some circumstances, the hardware will hand us a receive descriptor
which has no data attached, but is otherwise valid. The receive code was
improperly ignoring these descriptors, which result in an infinite loop.
To fix this, change the receive code to process all descriptors,
regardless of the size of the associated data. Add checks to the
memory-handling functions to allow for zero size.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_txrx.c | 15 ++++++++++++---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index c88e0701e1d7..e5c4c9139e54 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -607,6 +607,8 @@ ice_add_rx_frag(struct ice_rx_buf *rx_buf, struct sk_buff *skb,
unsigned int truesize = ICE_RXBUF_2048;
#endif
+ if (!size)
+ return;
skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_buf->page,
rx_buf->page_offset, size, truesize);
@@ -662,6 +664,8 @@ ice_get_rx_buf(struct ice_ring *rx_ring, struct sk_buff **skb,
prefetchw(rx_buf->page);
*skb = rx_buf->skb;
+ if (!size)
+ return rx_buf;
/* we are reusing so sync this buffer for CPU use */
dma_sync_single_range_for_cpu(rx_ring->dev, rx_buf->dma,
rx_buf->page_offset, size,
@@ -745,8 +749,11 @@ ice_construct_skb(struct ice_ring *rx_ring, struct ice_rx_buf *rx_buf,
*/
static void ice_put_rx_buf(struct ice_ring *rx_ring, struct ice_rx_buf *rx_buf)
{
- /* hand second half of page back to the ring */
+ if (!rx_buf)
+ return;
+
if (ice_can_reuse_rx_page(rx_buf)) {
+ /* hand second half of page back to the ring */
ice_reuse_rx_page(rx_ring, rx_buf);
rx_ring->rx_stats.page_reuse_count++;
} else {
@@ -1031,8 +1038,9 @@ static int ice_clean_rx_irq(struct ice_ring *rx_ring, int budget)
size = le16_to_cpu(rx_desc->wb.pkt_len) &
ICE_RX_FLX_DESC_PKT_LEN_M;
+ /* retrieve a buffer from the ring */
rx_buf = ice_get_rx_buf(rx_ring, &skb, size);
- /* allocate (if needed) and populate skb */
+
if (skb)
ice_add_rx_frag(rx_buf, skb, size);
else
@@ -1041,7 +1049,8 @@ static int ice_clean_rx_irq(struct ice_ring *rx_ring, int budget)
/* exit if we failed to retrieve a buffer */
if (!skb) {
rx_ring->rx_stats.alloc_buf_failed++;
- rx_buf->pagecnt_bias++;
+ if (rx_buf)
+ rx_buf->pagecnt_bias++;
break;
}
--
2.21.0
^ permalink raw reply related
* [net-next v2 03/14] ice: Restructure VFs initialization flows
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem
Cc: Akeem G Abodunrin, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
This patch restructures how VFs are configured, and resources allocated.
Instead of freeing resources that were never allocated, and resetting
empty VFs that have never been created - the new flow will just allocate
resources for number of requested VFs based on the availability.
During VFs initialization process, global interrupt is disabled, and
rearmed after getting MSIX vectors for VFs. This allows immediate mailbox
communications, instead of delaying it till later and VFs.
PF communications resulted to using polling instead of actual interrupt.
The issue manifested when creating higher number of VFs (128 VFs) per PF.
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice.h | 1 +
.../net/ethernet/intel/ice/ice_virtchnl_pf.c | 69 +++++++++++++------
2 files changed, 48 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 794d97460fc7..bde0b3617d9b 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -220,6 +220,7 @@ enum ice_state {
__ICE_CFG_BUSY,
__ICE_SERVICE_SCHED,
__ICE_SERVICE_DIS,
+ __ICE_OICR_INTR_DIS, /* Global OICR interrupt disabled */
__ICE_STATE_NBITS /* must be last */
};
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index ce01cbe70ea4..93b835a24bb1 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -974,6 +974,47 @@ ice_vf_set_vsi_promisc(struct ice_vf *vf, struct ice_vsi *vsi, u8 promisc_m,
return status;
}
+/**
+ * ice_config_res_vfs - Finalize allocation of VFs resources in one go
+ * @pf: pointer to the PF structure
+ *
+ * This function is being called as last part of resetting all VFs, or when
+ * configuring VFs for the first time, where there is no resource to be freed
+ * Returns true if resources were properly allocated for all VFs, and false
+ * otherwise.
+ */
+static bool ice_config_res_vfs(struct ice_pf *pf)
+{
+ struct ice_hw *hw = &pf->hw;
+ int v;
+
+ if (ice_check_avail_res(pf)) {
+ dev_err(&pf->pdev->dev,
+ "Cannot allocate VF resources, try with fewer number of VFs\n");
+ return false;
+ }
+
+ /* rearm global interrupts */
+ if (test_and_clear_bit(__ICE_OICR_INTR_DIS, pf->state))
+ ice_irq_dynamic_ena(hw, NULL, NULL);
+
+ /* Finish resetting each VF and allocate resources */
+ for (v = 0; v < pf->num_alloc_vfs; v++) {
+ struct ice_vf *vf = &pf->vf[v];
+
+ vf->num_vf_qs = pf->num_vf_qps;
+ dev_dbg(&pf->pdev->dev,
+ "VF-id %d has %d queues configured\n",
+ vf->vf_id, vf->num_vf_qs);
+ ice_cleanup_and_realloc_vf(vf);
+ }
+
+ ice_flush(hw);
+ clear_bit(__ICE_VF_DIS, pf->state);
+
+ return true;
+}
+
/**
* ice_reset_all_vfs - reset all allocated VFs in one go
* @pf: pointer to the PF structure
@@ -1066,25 +1107,8 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
dev_err(&pf->pdev->dev,
"Failed to free MSIX resources used by SR-IOV\n");
- if (ice_check_avail_res(pf)) {
- dev_err(&pf->pdev->dev,
- "Cannot allocate VF resources, try with fewer number of VFs\n");
+ if (!ice_config_res_vfs(pf))
return false;
- }
-
- /* Finish the reset on each VF */
- for (v = 0; v < pf->num_alloc_vfs; v++) {
- vf = &pf->vf[v];
-
- vf->num_vf_qs = pf->num_vf_qps;
- dev_dbg(&pf->pdev->dev,
- "VF-id %d has %d queues configured\n",
- vf->vf_id, vf->num_vf_qs);
- ice_cleanup_and_realloc_vf(vf);
- }
-
- ice_flush(hw);
- clear_bit(__ICE_VF_DIS, pf->state);
return true;
}
@@ -1249,7 +1273,7 @@ static int ice_alloc_vfs(struct ice_pf *pf, u16 num_alloc_vfs)
/* Disable global interrupt 0 so we don't try to handle the VFLR. */
wr32(hw, GLINT_DYN_CTL(pf->oicr_idx),
ICE_ITR_NONE << GLINT_DYN_CTL_ITR_INDX_S);
-
+ set_bit(__ICE_OICR_INTR_DIS, pf->state);
ice_flush(hw);
ret = pci_enable_sriov(pf->pdev, num_alloc_vfs);
@@ -1278,13 +1302,13 @@ static int ice_alloc_vfs(struct ice_pf *pf, u16 num_alloc_vfs)
}
pf->num_alloc_vfs = num_alloc_vfs;
- /* VF resources get allocated during reset */
- if (!ice_reset_all_vfs(pf, true)) {
+ /* VF resources get allocated with initialization */
+ if (!ice_config_res_vfs(pf)) {
ret = -EIO;
goto err_unroll_sriov;
}
- goto err_unroll_intr;
+ return ret;
err_unroll_sriov:
pf->vf = NULL;
@@ -1296,6 +1320,7 @@ static int ice_alloc_vfs(struct ice_pf *pf, u16 num_alloc_vfs)
err_unroll_intr:
/* rearm interrupts here */
ice_irq_dynamic_ena(hw, NULL, NULL);
+ clear_bit(__ICE_OICR_INTR_DIS, pf->state);
return ret;
}
--
2.21.0
^ permalink raw reply related
* [net-next v2 10/14] ice: Reduce wait times during VF bringup/reset
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem; +Cc: Brett Creeley, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
Currently there are a couple places where the VF is waiting too long when
checking the status of registers. This is causing the AVF driver to
spin for longer than necessary in the __IAVF_STARTUP state. Sometimes
it causes the AVF to go into the __IAVF_COMM_FAILED, which may retrigger
the __IAVF_STARTUP state. Try to reduce the chance of this happening by
removing unnecessary wait times in VF bringup/resets.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
.../net/ethernet/intel/ice/ice_virtchnl_pf.c | 26 +++++++++++--------
.../net/ethernet/intel/ice/ice_virtchnl_pf.h | 4 +++
2 files changed, 19 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 54b0e88af4ca..8a5bf9730fdf 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -382,12 +382,15 @@ static void ice_trigger_vf_reset(struct ice_vf *vf, bool is_vflr)
wr32(hw, PF_PCI_CIAA,
VF_DEVICE_STATUS | (vf_abs_id << PF_PCI_CIAA_VF_NUM_S));
- for (i = 0; i < 100; i++) {
+ for (i = 0; i < ICE_PCI_CIAD_WAIT_COUNT; i++) {
reg = rd32(hw, PF_PCI_CIAD);
- if ((reg & VF_TRANS_PENDING_M) != 0)
- dev_err(&pf->pdev->dev,
- "VF %d PCI transactions stuck\n", vf->vf_id);
- udelay(1);
+ /* no transactions pending so stop polling */
+ if ((reg & VF_TRANS_PENDING_M) == 0)
+ break;
+
+ dev_err(&pf->pdev->dev,
+ "VF %d PCI transactions stuck\n", vf->vf_id);
+ udelay(ICE_PCI_CIAD_WAIT_DELAY_US);
}
}
@@ -1068,7 +1071,6 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
* finished resetting.
*/
for (i = 0, v = 0; i < 10 && v < pf->num_alloc_vfs; i++) {
- usleep_range(10000, 20000);
/* Check each VF in sequence */
while (v < pf->num_alloc_vfs) {
@@ -1076,8 +1078,11 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
vf = &pf->vf[v];
reg = rd32(hw, VPGEN_VFRSTAT(vf->vf_id));
- if (!(reg & VPGEN_VFRSTAT_VFRD_M))
+ if (!(reg & VPGEN_VFRSTAT_VFRD_M)) {
+ /* only delay if the check failed */
+ usleep_range(10, 20);
break;
+ }
/* If the current VF has finished resetting, move on
* to the next VF in sequence.
@@ -1091,7 +1096,6 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr)
*/
if (v < pf->num_alloc_vfs)
dev_warn(&pf->pdev->dev, "VF reset check timeout\n");
- usleep_range(10000, 20000);
/* free VF resources to begin resetting the VSI state */
for (v = 0; v < pf->num_alloc_vfs; v++) {
@@ -1165,12 +1169,14 @@ static bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
* poll the status register to make sure that the reset
* completed successfully.
*/
- usleep_range(10000, 20000);
reg = rd32(hw, VPGEN_VFRSTAT(vf->vf_id));
if (reg & VPGEN_VFRSTAT_VFRD_M) {
rsd = true;
break;
}
+
+ /* only sleep if the reset is not done */
+ usleep_range(10, 20);
}
/* Display a warning if VF didn't manage to reset in time, but need to
@@ -1180,8 +1186,6 @@ static bool ice_reset_vf(struct ice_vf *vf, bool is_vflr)
dev_warn(&pf->pdev->dev, "VF reset check timeout on VF %d\n",
vf->vf_id);
- usleep_range(10000, 20000);
-
/* disable promiscuous modes in case they were enabled
* ignore any error if disabling process failed
*/
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
index 424bc0538956..79bb47f73879 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -22,6 +22,10 @@
#define VF_DEVICE_STATUS 0xAA
#define VF_TRANS_PENDING_M 0x20
+/* wait defines for polling PF_PCI_CIAD register status */
+#define ICE_PCI_CIAD_WAIT_COUNT 100
+#define ICE_PCI_CIAD_WAIT_DELAY_US 1
+
/* Specific VF states */
enum ice_vf_states {
ICE_VF_STATE_INIT = 0,
--
2.21.0
^ permalink raw reply related
* [net-next v2 14/14] ice: improve print for VF's when adding/deleting MAC filters
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem; +Cc: Brett Creeley, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
When we fail to add/delete MAC filters in the VF, the print doesn't
distinguish between the two. Fix that by printing whether or not we
failed to add/delete the MAC filter respectively.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 61cf1c5af928..1b1d1ea0c8f9 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -2283,8 +2283,8 @@ ice_vc_handle_mac_addr_msg(struct ice_vf *vf, u8 *msg, bool set)
if (v_ret) {
dev_err(&pf->pdev->dev,
- "can't update MAC filters for VF %d, error %d\n",
- vf->vf_id, v_ret);
+ "can't %s MAC filters for VF %d, error %d\n",
+ set ? "add" : "remove", vf->vf_id, v_ret);
} else {
if (set)
vf->num_mac += mac_count;
--
2.21.0
^ permalink raw reply related
* [net-next v2 05/14] ice: Set WB_ON_ITR when we don't re-enable interrupts
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem
Cc: Brett Creeley, netdev, nhorman, sassmann, Tony Nguyen,
Andrew Bowers, Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
Currently when busy polling is enabled we aren't setting/enabling
WB_ON_ITR in the driver. This doesn't break the driver, but it does
cause issues. If we don't enable WB_ON_ITR mode we will still get
write-backs from hardware during polling when a cache line has been
filled, but if a cache line is not filled we will not get the
write-back because WB_ON_ITR is not set. Fix this by enabling
WB_ON_ITR in the driver when interrupts are disabled.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
.../net/ethernet/intel/ice/ice_hw_autogen.h | 3 ++
drivers/net/ethernet/intel/ice/ice_txrx.c | 54 +++++++++++++++++++
drivers/net/ethernet/intel/ice/ice_txrx.h | 13 +++++
3 files changed, 70 insertions(+)
diff --git a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
index 87652d722a30..6f78ff5534af 100644
--- a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
+++ b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
@@ -127,8 +127,11 @@
#define GLINT_DYN_CTL_CLEARPBA_M BIT(1)
#define GLINT_DYN_CTL_SWINT_TRIG_M BIT(2)
#define GLINT_DYN_CTL_ITR_INDX_S 3
+#define GLINT_DYN_CTL_ITR_INDX_M ICE_M(0x3, 3)
#define GLINT_DYN_CTL_INTERVAL_S 5
+#define GLINT_DYN_CTL_INTERVAL_M ICE_M(0xFFF, 5)
#define GLINT_DYN_CTL_SW_ITR_INDX_M ICE_M(0x3, 25)
+#define GLINT_DYN_CTL_WB_ON_ITR_M BIT(30)
#define GLINT_DYN_CTL_INTENA_MSK_M BIT(31)
#define GLINT_ITR(_i, _INT) (0x00154000 + ((_i) * 8192 + (_INT) * 4))
#define GLINT_RATE(_INT) (0x0015A000 + ((_INT) * 4))
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index 837d6ae2f33b..c88e0701e1d7 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -1355,6 +1355,23 @@ ice_update_ena_itr(struct ice_vsi *vsi, struct ice_q_vector *q_vector)
struct ice_ring_container *rx = &q_vector->rx;
u32 itr_val;
+ /* when exiting WB_ON_ITR lets set a low ITR value and trigger
+ * interrupts to expire right away in case we have more work ready to go
+ * already
+ */
+ if (q_vector->itr_countdown == ICE_IN_WB_ON_ITR_MODE) {
+ itr_val = ice_buildreg_itr(rx->itr_idx, ICE_WB_ON_ITR_USECS);
+ wr32(&vsi->back->hw, GLINT_DYN_CTL(q_vector->reg_idx), itr_val);
+ /* set target back to last user set value */
+ rx->target_itr = rx->itr_setting;
+ /* set current to what we just wrote and dynamic if needed */
+ rx->current_itr = ICE_WB_ON_ITR_USECS |
+ (rx->itr_setting & ICE_ITR_DYNAMIC);
+ /* allow normal interrupt flow to start */
+ q_vector->itr_countdown = 0;
+ return;
+ }
+
/* This will do nothing if dynamic updates are not enabled */
ice_update_itr(q_vector, tx);
ice_update_itr(q_vector, rx);
@@ -1399,6 +1416,41 @@ ice_update_ena_itr(struct ice_vsi *vsi, struct ice_q_vector *q_vector)
itr_val);
}
+/**
+ * ice_set_wb_on_itr - set WB_ON_ITR for this q_vector
+ * @vsi: pointer to the VSI structure
+ * @q_vector: q_vector to set WB_ON_ITR on
+ *
+ * We need to tell hardware to write-back completed descriptors even when
+ * interrupts are disabled. Descriptors will be written back on cache line
+ * boundaries without WB_ON_ITR enabled, but if we don't enable WB_ON_ITR
+ * descriptors may not be written back if they don't fill a cache line until the
+ * next interrupt.
+ *
+ * This sets the write-back frequency to 2 microseconds as that is the minimum
+ * value that's not 0 due to ITR granularity. Also, set the INTENA_MSK bit to
+ * make sure hardware knows we aren't meddling with the INTENA_M bit.
+ */
+static void
+ice_set_wb_on_itr(struct ice_vsi *vsi, struct ice_q_vector *q_vector)
+{
+ /* already in WB_ON_ITR mode no need to change it */
+ if (q_vector->itr_countdown == ICE_IN_WB_ON_ITR_MODE)
+ return;
+
+ if (q_vector->num_ring_rx)
+ wr32(&vsi->back->hw, GLINT_DYN_CTL(q_vector->reg_idx),
+ ICE_GLINT_DYN_CTL_WB_ON_ITR(ICE_WB_ON_ITR_USECS,
+ ICE_RX_ITR));
+
+ if (q_vector->num_ring_tx)
+ wr32(&vsi->back->hw, GLINT_DYN_CTL(q_vector->reg_idx),
+ ICE_GLINT_DYN_CTL_WB_ON_ITR(ICE_WB_ON_ITR_USECS,
+ ICE_TX_ITR));
+
+ q_vector->itr_countdown = ICE_IN_WB_ON_ITR_MODE;
+}
+
/**
* ice_napi_poll - NAPI polling Rx/Tx cleanup routine
* @napi: napi struct with our devices info in it
@@ -1459,6 +1511,8 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
*/
if (likely(napi_complete_done(napi, work_done)))
ice_update_ena_itr(vsi, q_vector);
+ else
+ ice_set_wb_on_itr(vsi, q_vector);
return min_t(int, work_done, budget - 1);
}
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.h b/drivers/net/ethernet/intel/ice/ice_txrx.h
index ec76aba347b9..94a9280193e2 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.h
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.h
@@ -144,6 +144,19 @@ enum ice_rx_dtype {
#define ICE_DFLT_INTRL 0
#define ICE_MAX_INTRL 236
+#define ICE_WB_ON_ITR_USECS 2
+#define ICE_IN_WB_ON_ITR_MODE 255
+/* Sets WB_ON_ITR and assumes INTENA bit is already cleared, which allows
+ * setting the MSK_M bit to tell hardware to ignore the INTENA_M bit. Also,
+ * set the write-back latency to the usecs passed in.
+ */
+#define ICE_GLINT_DYN_CTL_WB_ON_ITR(usecs, itr_idx) \
+ ((((usecs) << (GLINT_DYN_CTL_INTERVAL_S - ICE_ITR_GRAN_S)) & \
+ GLINT_DYN_CTL_INTERVAL_M) | \
+ (((itr_idx) << GLINT_DYN_CTL_ITR_INDX_S) & \
+ GLINT_DYN_CTL_ITR_INDX_M) | GLINT_DYN_CTL_INTENA_MSK_M | \
+ GLINT_DYN_CTL_WB_ON_ITR_M)
+
/* Legacy or Advanced Mode Queue */
#define ICE_TX_ADVANCED 0
#define ICE_TX_LEGACY 1
--
2.21.0
^ permalink raw reply related
* [net-next v2 11/14] ice: Increase size of Mailbox receive queue for many VFs
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem; +Cc: Brett Creeley, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
Currently we use the ICE_MBXQ_LEN for both the Mailbox send and receive
queues that are used to communicate with VFs. This is fine for the send
queue because the PF driver will lock the queue for every single send,
but for the Mailbox receive queue every VF is posting to its Mailbox
send queue and the hardware is then handing the message to the PF on its
Mailbox receive queue. This becomes a problem with many VFs because it
seems to overburden the Mailbox receive queue on the PF. Fix this by
increasing the Mailbox receive queue for the PF to 512 entries.
The number 512 was determined based on the number of VFs supported by
the device. We can have a total of 256 VFs so in the worst case this
allows the VFs to put 2 messages in the PFs Mailbox receive queue at the
same time.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice.h | 3 ++-
drivers/net/ethernet/intel/ice/ice_main.c | 4 ++--
2 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index bde0b3617d9b..b60e64c0036b 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -69,7 +69,8 @@ extern const char ice_drv_ver[];
#define ICE_INT_NAME_STR_LEN (IFNAMSIZ + 16)
#define ICE_ETHTOOL_FWVER_LEN 32
#define ICE_AQ_LEN 64
-#define ICE_MBXQ_LEN 64
+#define ICE_MBXSQ_LEN 64
+#define ICE_MBXRQ_LEN 512
#define ICE_MIN_MSIX 2
#define ICE_NO_VSI 0xffff
#define ICE_MAX_TXQS 2048
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 7805c2abd4ac..a0d148f590c2 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1507,8 +1507,8 @@ static void ice_set_ctrlq_len(struct ice_hw *hw)
hw->adminq.num_sq_entries = ICE_AQ_LEN;
hw->adminq.rq_buf_size = ICE_AQ_MAX_BUF_LEN;
hw->adminq.sq_buf_size = ICE_AQ_MAX_BUF_LEN;
- hw->mailboxq.num_rq_entries = ICE_MBXQ_LEN;
- hw->mailboxq.num_sq_entries = ICE_MBXQ_LEN;
+ hw->mailboxq.num_rq_entries = ICE_MBXRQ_LEN;
+ hw->mailboxq.num_sq_entries = ICE_MBXSQ_LEN;
hw->mailboxq.rq_buf_size = ICE_MBXQ_MAX_BUF_LEN;
hw->mailboxq.sq_buf_size = ICE_MBXQ_MAX_BUF_LEN;
}
--
2.21.0
^ permalink raw reply related
* [net-next v2 13/14] ice: Change type for queue counts
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem
Cc: Pawel Kaminski, netdev, nhorman, sassmann, Tony Nguyen,
Andrew Bowers, Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Pawel Kaminski <pawel.kaminski@intel.com>
These queue variables are being assigned values that are type u16.
Change the local variables to match these types. Since these
represent queue counts, they should never be negative.
Signed-off-by: Pawel Kaminski <pawel.kaminski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
.../net/ethernet/intel/ice/ice_virtchnl_pf.c | 25 ++++++++++---------
1 file changed, 13 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 8a5bf9730fdf..61cf1c5af928 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -2337,11 +2337,11 @@ static int ice_vc_request_qs_msg(struct ice_vf *vf, u8 *msg)
enum virtchnl_status_code v_ret = VIRTCHNL_STATUS_SUCCESS;
struct virtchnl_vf_res_request *vfres =
(struct virtchnl_vf_res_request *)msg;
- int req_queues = vfres->num_queue_pairs;
+ u16 req_queues = vfres->num_queue_pairs;
struct ice_pf *pf = vf->pf;
- int max_allowed_vf_queues;
- int tx_rx_queue_left;
- int cur_queues;
+ u16 max_allowed_vf_queues;
+ u16 tx_rx_queue_left;
+ u16 cur_queues;
if (!test_bit(ICE_VF_STATE_ACTIVE, vf->vf_states)) {
v_ret = VIRTCHNL_STATUS_ERR_PARAM;
@@ -2349,29 +2349,30 @@ static int ice_vc_request_qs_msg(struct ice_vf *vf, u8 *msg)
}
cur_queues = vf->num_vf_qs;
- tx_rx_queue_left = min_t(int, pf->q_left_tx, pf->q_left_rx);
+ tx_rx_queue_left = min_t(u16, pf->q_left_tx, pf->q_left_rx);
max_allowed_vf_queues = tx_rx_queue_left + cur_queues;
- if (req_queues <= 0) {
+ if (!req_queues) {
dev_err(&pf->pdev->dev,
- "VF %d tried to request %d queues. Ignoring.\n",
- vf->vf_id, req_queues);
+ "VF %d tried to request 0 queues. Ignoring.\n",
+ vf->vf_id);
} else if (req_queues > ICE_MAX_BASE_QS_PER_VF) {
dev_err(&pf->pdev->dev,
"VF %d tried to request more than %d queues.\n",
vf->vf_id, ICE_MAX_BASE_QS_PER_VF);
vfres->num_queue_pairs = ICE_MAX_BASE_QS_PER_VF;
- } else if (req_queues - cur_queues > tx_rx_queue_left) {
+ } else if (req_queues > cur_queues &&
+ req_queues - cur_queues > tx_rx_queue_left) {
dev_warn(&pf->pdev->dev,
- "VF %d requested %d more queues, but only %d left.\n",
+ "VF %d requested %u more queues, but only %u left.\n",
vf->vf_id, req_queues - cur_queues, tx_rx_queue_left);
- vfres->num_queue_pairs = min_t(int, max_allowed_vf_queues,
+ vfres->num_queue_pairs = min_t(u16, max_allowed_vf_queues,
ICE_MAX_BASE_QS_PER_VF);
} else {
/* request is successful, then reset VF */
vf->num_req_qs = req_queues;
ice_vc_dis_vf(vf);
dev_info(&pf->pdev->dev,
- "VF %d granted request of %d queues.\n",
+ "VF %d granted request of %u queues.\n",
vf->vf_id, req_queues);
return 0;
}
--
2.21.0
^ permalink raw reply related
* [net-next v2 04/14] ice: fix set pause param autoneg check
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem
Cc: Paul Greenwalt, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Paul Greenwalt <paul.greenwalt@intel.com>
When ETHTOOL_GLINKSETTINGS is defined get pause param pause->autoneg
reports SW configured setting, however when not defined get pause param
pause->autoneg reports the link status. Set pause param needs to compare
pause->autoneg with the same source as get pause param to block the user
from changing autoneg with the set pause param option, or the user
may be incorrectly blocked from changing Rx|Tx pause settings.
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_ethtool.c | 28 +++++++++++++++++++-
1 file changed, 27 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index d3ba535bd65a..d02b72e72423 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -2856,6 +2856,7 @@ static int
ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
+ struct ice_aqc_get_phy_caps_data *pcaps;
struct ice_link_status *hw_link_info;
struct ice_pf *pf = np->vsi->back;
struct ice_dcbx_cfg *dcbx_cfg;
@@ -2866,6 +2867,7 @@ ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
u8 aq_failures;
bool link_up;
int err = 0;
+ u32 is_an;
pi = vsi->port_info;
hw_link_info = &pi->phy.link_info;
@@ -2880,7 +2882,31 @@ ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
return -EOPNOTSUPP;
}
- if (pause->autoneg != (hw_link_info->an_info & ICE_AQ_AN_COMPLETED)) {
+ /* Get pause param reports configured and negotiated flow control pause
+ * when ETHTOOL_GLINKSETTINGS is defined. Since ETHTOOL_GLINKSETTINGS is
+ * defined get pause param pause->autoneg reports SW configured setting,
+ * so compare pause->autoneg with SW configured to prevent the user from
+ * using set pause param to chance autoneg.
+ */
+ pcaps = devm_kzalloc(&vsi->back->pdev->dev, sizeof(*pcaps),
+ GFP_KERNEL);
+ if (!pcaps)
+ return -ENOMEM;
+
+ /* Get current PHY config */
+ status = ice_aq_get_phy_caps(pi, false, ICE_AQC_REPORT_SW_CFG, pcaps,
+ NULL);
+ if (status) {
+ devm_kfree(&vsi->back->pdev->dev, pcaps);
+ return -EIO;
+ }
+
+ is_an = ((pcaps->caps & ICE_AQC_PHY_AN_MODE) ?
+ AUTONEG_ENABLE : AUTONEG_DISABLE);
+
+ devm_kfree(&vsi->back->pdev->dev, pcaps);
+
+ if (pause->autoneg != is_an) {
netdev_info(netdev, "To change autoneg please use: ethtool -s <dev> autoneg <on|off>\n");
return -EOPNOTSUPP;
}
--
2.21.0
^ permalink raw reply related
* [net-next v2 12/14] ice: Move VF resources definition to SR-IOV specific file
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem
Cc: Akeem G Abodunrin, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
In order to use some of the VF resources definition in the SR-IOV specific
virtchnl header file, this patch moves applicable code to
ice_virtchnl_pf.h file accordingly... and they should have been defined in
the destination file originally.
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice.h | 10 ----------
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h | 13 +++++++++++++
2 files changed, 13 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index b60e64c0036b..9f9c30d29eb5 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -87,16 +87,6 @@ extern const char ice_drv_ver[];
#define ICE_RES_MISC_VEC_ID (ICE_RES_VALID_BIT - 1)
#define ICE_INVAL_Q_INDEX 0xffff
#define ICE_INVAL_VFID 256
-#define ICE_MAX_VF_COUNT 256
-#define ICE_MAX_QS_PER_VF 256
-#define ICE_MIN_QS_PER_VF 1
-#define ICE_DFLT_QS_PER_VF 4
-#define ICE_NONQ_VECS_VF 1
-#define ICE_MAX_SCATTER_QS_PER_VF 16
-#define ICE_MAX_BASE_QS_PER_VF 16
-#define ICE_MAX_INTR_PER_VF 65
-#define ICE_MIN_INTR_PER_VF (ICE_MIN_QS_PER_VF + 1)
-#define ICE_DFLT_INTR_PER_VF (ICE_DFLT_QS_PER_VF + 1)
#define ICE_MAX_RESET_WAIT 20
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
index 79bb47f73879..4d94853f119a 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -26,6 +26,19 @@
#define ICE_PCI_CIAD_WAIT_COUNT 100
#define ICE_PCI_CIAD_WAIT_DELAY_US 1
+/* VF resources default values and limitation */
+#define ICE_MAX_VF_COUNT 256
+#define ICE_MAX_QS_PER_VF 256
+#define ICE_MIN_QS_PER_VF 1
+#define ICE_DFLT_QS_PER_VF 4
+#define ICE_NONQ_VECS_VF 1
+#define ICE_MAX_SCATTER_QS_PER_VF 16
+#define ICE_MAX_BASE_QS_PER_VF 16
+#define ICE_MAX_INTR_PER_VF 65
+#define ICE_MAX_POLICY_INTR_PER_VF 33
+#define ICE_MIN_INTR_PER_VF (ICE_MIN_QS_PER_VF + 1)
+#define ICE_DFLT_INTR_PER_VF (ICE_DFLT_QS_PER_VF + 1)
+
/* Specific VF states */
enum ice_vf_states {
ICE_VF_STATE_INIT = 0,
--
2.21.0
^ permalink raw reply related
* [net-next v2 09/14] ice: update GLINT_DYN_CTL and GLINT_VECT2FUNC register access
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem
Cc: Paul Greenwalt, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Paul Greenwalt <paul.greenwalt@intel.com>
Register access for GLINT_DYN_CTL and GLINT_VECT2FUNC should be within
the PF space and not the absolute device space.
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
.../net/ethernet/intel/ice/ice_virtchnl_pf.c | 24 +++++++++++--------
.../net/ethernet/intel/ice/ice_virtchnl_pf.h | 3 ++-
2 files changed, 16 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 93b835a24bb1..54b0e88af4ca 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -474,19 +474,20 @@ ice_vf_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi, u16 vf_id)
}
/**
- * ice_calc_vf_first_vector_idx - Calculate absolute MSIX vector index in HW
+ * ice_calc_vf_first_vector_idx - Calculate MSIX vector index in the PF space
* @pf: pointer to PF structure
* @vf: pointer to VF that the first MSIX vector index is being calculated for
*
- * This returns the first MSIX vector index in HW that is used by this VF and
- * this will always be the OICR index in the AVF driver so any functionality
+ * This returns the first MSIX vector index in PF space that is used by this VF.
+ * This index is used when accessing PF relative registers such as
+ * GLINT_VECT2FUNC and GLINT_DYN_CTL.
+ * This will always be the OICR index in the AVF driver so any functionality
* using vf->first_vector_idx for queue configuration will have to increment by
* 1 to avoid meddling with the OICR index.
*/
static int ice_calc_vf_first_vector_idx(struct ice_pf *pf, struct ice_vf *vf)
{
- return pf->hw.func_caps.common_cap.msix_vector_first_id +
- pf->sriov_base_vector + vf->vf_id * pf->num_vf_msix;
+ return pf->sriov_base_vector + vf->vf_id * pf->num_vf_msix;
}
/**
@@ -597,27 +598,30 @@ static int ice_alloc_vf_res(struct ice_vf *vf)
*/
static void ice_ena_vf_mappings(struct ice_vf *vf)
{
+ int abs_vf_id, abs_first, abs_last;
struct ice_pf *pf = vf->pf;
struct ice_vsi *vsi;
int first, last, v;
struct ice_hw *hw;
- int abs_vf_id;
u32 reg;
hw = &pf->hw;
vsi = pf->vsi[vf->lan_vsi_idx];
first = vf->first_vector_idx;
last = (first + pf->num_vf_msix) - 1;
+ abs_first = first + pf->hw.func_caps.common_cap.msix_vector_first_id;
+ abs_last = (abs_first + pf->num_vf_msix) - 1;
abs_vf_id = vf->vf_id + hw->func_caps.vf_base_id;
/* VF Vector allocation */
- reg = (((first << VPINT_ALLOC_FIRST_S) & VPINT_ALLOC_FIRST_M) |
- ((last << VPINT_ALLOC_LAST_S) & VPINT_ALLOC_LAST_M) |
+ reg = (((abs_first << VPINT_ALLOC_FIRST_S) & VPINT_ALLOC_FIRST_M) |
+ ((abs_last << VPINT_ALLOC_LAST_S) & VPINT_ALLOC_LAST_M) |
VPINT_ALLOC_VALID_M);
wr32(hw, VPINT_ALLOC(vf->vf_id), reg);
- reg = (((first << VPINT_ALLOC_PCI_FIRST_S) & VPINT_ALLOC_PCI_FIRST_M) |
- ((last << VPINT_ALLOC_PCI_LAST_S) & VPINT_ALLOC_PCI_LAST_M) |
+ reg = (((abs_first << VPINT_ALLOC_PCI_FIRST_S)
+ & VPINT_ALLOC_PCI_FIRST_M) |
+ ((abs_last << VPINT_ALLOC_PCI_LAST_S) & VPINT_ALLOC_PCI_LAST_M) |
VPINT_ALLOC_PCI_VALID_M);
wr32(hw, VPINT_ALLOC_PCI(vf->vf_id), reg);
/* map the interrupts to its functions */
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
index ada69120ff38..424bc0538956 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h
@@ -45,7 +45,8 @@ struct ice_vf {
s16 vf_id; /* VF ID in the PF space */
u16 lan_vsi_idx; /* index into PF struct */
- int first_vector_idx; /* first vector index of this VF */
+ /* first vector index of this VF in the PF space */
+ int first_vector_idx;
struct ice_sw *vf_sw_id; /* switch ID the VF VSIs connect to */
struct virtchnl_version_info vf_ver;
u32 driver_caps; /* reported by VF driver */
--
2.21.0
^ permalink raw reply related
* [net-next v2 01/14] ice: Use the software based tail when checking for hung Tx ring
From: Jeff Kirsher @ 2019-08-19 16:16 UTC (permalink / raw)
To: davem; +Cc: Brett Creeley, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Brett Creeley <brett.creeley@intel.com>
Currently in ice_get_tx_pending we try to read a Tx ring's tail. This is
then compared with the software based head (next_to_clean) to determine
if we have pending work. This will never work because reading of the Tx
ring's tail is no longer supported. Fix this by using the software based
tail (next_to_use) to determine if there is pending work.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_main.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index c26e6a102dac..1aa7e06ebbdc 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -41,12 +41,12 @@ static void ice_update_pf_stats(struct ice_pf *pf);
* ice_get_tx_pending - returns number of Tx descriptors not processed
* @ring: the ring of descriptors
*/
-static u32 ice_get_tx_pending(struct ice_ring *ring)
+static u16 ice_get_tx_pending(struct ice_ring *ring)
{
- u32 head, tail;
+ u16 head, tail;
head = ring->next_to_clean;
- tail = readl(ring->tail);
+ tail = ring->next_to_use;
if (head != tail)
return (head < tail) ?
--
2.21.0
^ permalink raw reply related
* [net-next v2 06/14] ice: Fix kernel hang with DCB reset in CEE mode
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem; +Cc: Usha Ketineni, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Usha Ketineni <usha.k.ketineni@intel.com>
This patch fixes the set local MIB AQ call failures in the DCB rebuild path
by setting the defaults for the ETS recommended DCB configuration. Also,
willing bits for the DCB configuration needs to be set correctly. Resets
works fine in IEEE mode as the ETS recommended DCB configuration is
populated but not in CEE mode.
Without this patch, PFR causes the kernel hang in CEE mode.
Signed-off-by: Usha Ketineni <usha.k.ketineni@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_dcb_lib.c | 149 +++++++++++--------
1 file changed, 88 insertions(+), 61 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_dcb_lib.c b/drivers/net/ethernet/intel/ice/ice_dcb_lib.c
index fe88b127ca42..bf6cd4760a48 100644
--- a/drivers/net/ethernet/intel/ice/ice_dcb_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_dcb_lib.c
@@ -203,16 +203,87 @@ int ice_pf_dcb_cfg(struct ice_pf *pf, struct ice_dcbx_cfg *new_cfg, bool locked)
return ret;
}
+/**
+ * ice_cfg_etsrec_defaults - Set default ETS recommended DCB config
+ * @pi: port information structure
+ */
+static void ice_cfg_etsrec_defaults(struct ice_port_info *pi)
+{
+ struct ice_dcbx_cfg *dcbcfg = &pi->local_dcbx_cfg;
+ u8 i;
+
+ /* Ensure ETS recommended DCB configuration is not already set */
+ if (dcbcfg->etsrec.maxtcs)
+ return;
+
+ /* In CEE mode, set the default to 1 TC */
+ dcbcfg->etsrec.maxtcs = 1;
+ for (i = 0; i < ICE_MAX_TRAFFIC_CLASS; i++) {
+ dcbcfg->etsrec.tcbwtable[i] = i ? 0 : 100;
+ dcbcfg->etsrec.tsatable[i] = i ? ICE_IEEE_TSA_STRICT :
+ ICE_IEEE_TSA_ETS;
+ }
+}
+
+/**
+ * ice_dcb_need_recfg - Check if DCB needs reconfig
+ * @pf: board private structure
+ * @old_cfg: current DCB config
+ * @new_cfg: new DCB config
+ */
+static bool
+ice_dcb_need_recfg(struct ice_pf *pf, struct ice_dcbx_cfg *old_cfg,
+ struct ice_dcbx_cfg *new_cfg)
+{
+ bool need_reconfig = false;
+
+ /* Check if ETS configuration has changed */
+ if (memcmp(&new_cfg->etscfg, &old_cfg->etscfg,
+ sizeof(new_cfg->etscfg))) {
+ /* If Priority Table has changed reconfig is needed */
+ if (memcmp(&new_cfg->etscfg.prio_table,
+ &old_cfg->etscfg.prio_table,
+ sizeof(new_cfg->etscfg.prio_table))) {
+ need_reconfig = true;
+ dev_dbg(&pf->pdev->dev, "ETS UP2TC changed.\n");
+ }
+
+ if (memcmp(&new_cfg->etscfg.tcbwtable,
+ &old_cfg->etscfg.tcbwtable,
+ sizeof(new_cfg->etscfg.tcbwtable)))
+ dev_dbg(&pf->pdev->dev, "ETS TC BW Table changed.\n");
+
+ if (memcmp(&new_cfg->etscfg.tsatable,
+ &old_cfg->etscfg.tsatable,
+ sizeof(new_cfg->etscfg.tsatable)))
+ dev_dbg(&pf->pdev->dev, "ETS TSA Table changed.\n");
+ }
+
+ /* Check if PFC configuration has changed */
+ if (memcmp(&new_cfg->pfc, &old_cfg->pfc, sizeof(new_cfg->pfc))) {
+ need_reconfig = true;
+ dev_dbg(&pf->pdev->dev, "PFC config change detected.\n");
+ }
+
+ /* Check if APP Table has changed */
+ if (memcmp(&new_cfg->app, &old_cfg->app, sizeof(new_cfg->app))) {
+ need_reconfig = true;
+ dev_dbg(&pf->pdev->dev, "APP Table change detected.\n");
+ }
+
+ dev_dbg(&pf->pdev->dev, "dcb need_reconfig=%d\n", need_reconfig);
+ return need_reconfig;
+}
+
/**
* ice_dcb_rebuild - rebuild DCB post reset
* @pf: physical function instance
*/
void ice_dcb_rebuild(struct ice_pf *pf)
{
+ struct ice_dcbx_cfg *local_dcbx_cfg, *desired_dcbx_cfg, *prev_cfg;
struct ice_aqc_port_ets_elem buf = { 0 };
- struct ice_dcbx_cfg *prev_cfg;
enum ice_status ret;
- u8 willing;
ret = ice_query_port_ets(pf->hw.port_info, &buf, sizeof(buf), NULL);
if (ret) {
@@ -224,9 +295,15 @@ void ice_dcb_rebuild(struct ice_pf *pf)
if (!test_bit(ICE_FLAG_DCB_ENA, pf->flags))
return;
+ local_dcbx_cfg = &pf->hw.port_info->local_dcbx_cfg;
+ desired_dcbx_cfg = &pf->hw.port_info->desired_dcbx_cfg;
+
/* Save current willing state and force FW to unwilling */
- willing = pf->hw.port_info->local_dcbx_cfg.etscfg.willing;
- pf->hw.port_info->local_dcbx_cfg.etscfg.willing = 0x0;
+ local_dcbx_cfg->etscfg.willing = 0x0;
+ local_dcbx_cfg->pfc.willing = 0x0;
+ local_dcbx_cfg->app_mode = ICE_DCBX_APPS_NON_WILLING;
+
+ ice_cfg_etsrec_defaults(pf->hw.port_info);
ret = ice_set_dcb_cfg(pf->hw.port_info);
if (ret) {
dev_err(&pf->pdev->dev, "Failed to set DCB to unwilling\n");
@@ -234,8 +311,7 @@ void ice_dcb_rebuild(struct ice_pf *pf)
}
/* Retrieve DCB config and ensure same as current in SW */
- prev_cfg = devm_kmemdup(&pf->pdev->dev,
- &pf->hw.port_info->local_dcbx_cfg,
+ prev_cfg = devm_kmemdup(&pf->pdev->dev, local_dcbx_cfg,
sizeof(*prev_cfg), GFP_KERNEL);
if (!prev_cfg) {
dev_err(&pf->pdev->dev, "Failed to alloc space for DCB cfg\n");
@@ -243,22 +319,22 @@ void ice_dcb_rebuild(struct ice_pf *pf)
}
ice_init_dcb(&pf->hw);
- if (memcmp(prev_cfg, &pf->hw.port_info->local_dcbx_cfg,
- sizeof(*prev_cfg))) {
+ if (ice_dcb_need_recfg(pf, prev_cfg, local_dcbx_cfg)) {
/* difference in cfg detected - disable DCB till next MIB */
dev_err(&pf->pdev->dev, "Set local MIB not accurate\n");
- devm_kfree(&pf->pdev->dev, prev_cfg);
goto dcb_error;
}
/* fetched config congruent to previous configuration */
devm_kfree(&pf->pdev->dev, prev_cfg);
- /* Configuration replayed - reset willing state to previous */
- pf->hw.port_info->local_dcbx_cfg.etscfg.willing = willing;
+ /* Set the local desired config */
+ memset(&pf->hw.port_info->local_dcbx_cfg, 0, sizeof(*local_dcbx_cfg));
+ memcpy(local_dcbx_cfg, desired_dcbx_cfg, sizeof(*local_dcbx_cfg));
+ ice_cfg_etsrec_defaults(pf->hw.port_info);
ret = ice_set_dcb_cfg(pf->hw.port_info);
if (ret) {
- dev_err(&pf->pdev->dev, "Fail restoring prev willing state\n");
+ dev_err(&pf->pdev->dev, "Failed to set desired config\n");
goto dcb_error;
}
dev_info(&pf->pdev->dev, "DCB restored after reset\n");
@@ -501,55 +577,6 @@ ice_tx_prepare_vlan_flags_dcb(struct ice_ring *tx_ring,
return 0;
}
-/**
- * ice_dcb_need_recfg - Check if DCB needs reconfig
- * @pf: board private structure
- * @old_cfg: current DCB config
- * @new_cfg: new DCB config
- */
-static bool ice_dcb_need_recfg(struct ice_pf *pf, struct ice_dcbx_cfg *old_cfg,
- struct ice_dcbx_cfg *new_cfg)
-{
- bool need_reconfig = false;
-
- /* Check if ETS configuration has changed */
- if (memcmp(&new_cfg->etscfg, &old_cfg->etscfg,
- sizeof(new_cfg->etscfg))) {
- /* If Priority Table has changed reconfig is needed */
- if (memcmp(&new_cfg->etscfg.prio_table,
- &old_cfg->etscfg.prio_table,
- sizeof(new_cfg->etscfg.prio_table))) {
- need_reconfig = true;
- dev_dbg(&pf->pdev->dev, "ETS UP2TC changed.\n");
- }
-
- if (memcmp(&new_cfg->etscfg.tcbwtable,
- &old_cfg->etscfg.tcbwtable,
- sizeof(new_cfg->etscfg.tcbwtable)))
- dev_dbg(&pf->pdev->dev, "ETS TC BW Table changed.\n");
-
- if (memcmp(&new_cfg->etscfg.tsatable,
- &old_cfg->etscfg.tsatable,
- sizeof(new_cfg->etscfg.tsatable)))
- dev_dbg(&pf->pdev->dev, "ETS TSA Table changed.\n");
- }
-
- /* Check if PFC configuration has changed */
- if (memcmp(&new_cfg->pfc, &old_cfg->pfc, sizeof(new_cfg->pfc))) {
- need_reconfig = true;
- dev_dbg(&pf->pdev->dev, "PFC config change detected.\n");
- }
-
- /* Check if APP Table has changed */
- if (memcmp(&new_cfg->app, &old_cfg->app, sizeof(new_cfg->app))) {
- need_reconfig = true;
- dev_dbg(&pf->pdev->dev, "APP Table change detected.\n");
- }
-
- dev_dbg(&pf->pdev->dev, "dcb need_reconfig=%d\n", need_reconfig);
- return need_reconfig;
-}
-
/**
* ice_dcb_process_lldp_set_mib_change - Process MIB change
* @pf: ptr to ice_pf
--
2.21.0
^ permalink raw reply related
* [net-next v2 08/14] ice: Do not always bring up PF VSI in ice_ena_vsi()
From: Jeff Kirsher @ 2019-08-19 16:17 UTC (permalink / raw)
To: davem; +Cc: Tony Nguyen, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190819161708.3763-1-jeffrey.t.kirsher@intel.com>
From: Tony Nguyen <anthony.l.nguyen@intel.com>
During rebuild ice_ena_vsi() is called to recover the VSI state.
This function assumes the PF VSI is always to be enabled, however,
it's possible that during reset/rebuild the interface can be
brought down. If this occurs, we can attempt to bring up the PF
VSI on a downed interface which can lead to various crashes. If
the interface is not running, do not bring up the associated VSI.
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ice/ice_main.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 1aa7e06ebbdc..7805c2abd4ac 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -3701,8 +3701,6 @@ static int ice_ena_vsi(struct ice_vsi *vsi, bool locked)
err = netd->netdev_ops->ndo_open(netd);
rtnl_unlock();
}
- } else {
- err = ice_vsi_open(vsi);
}
}
--
2.21.0
^ permalink raw reply related
* Re: [PATCH net-next 4/6] net: dsa: mv88e6xxx: do not change STP state on port disabling
From: Vivien Didelot @ 2019-08-19 16:27 UTC (permalink / raw)
To: Andrew Lunn; +Cc: netdev, marek.behun, davem, f.fainelli
In-Reply-To: <20190819161018.GI15291@lunn.ch>
On Mon, 19 Aug 2019 18:10:18 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> > On Mon, 19 Aug 2019 15:40:57 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> > > On Sun, Aug 18, 2019 at 01:35:46PM -0400, Vivien Didelot wrote:
> > > > When disabling a port, that is not for the driver to decide what to
> > > > do with the STP state. This is already handled by the DSA layer.
> > >
> > > Putting the port into STP disabled state is how you actually disable
> > > it, for the mv88e6xxx. So this is not really about STP, it is about
> > > powering off the port. Maybe a comment is needed, rather than removing
> > > the code?
> >
> > This is not for the driver to decide, the stack already handles that.
> > Otherwise, calling dsa_port_disable on a bridged port would result in
> > mv88e6xxx forcing the STP state to Disabled while this is not expected.
[...]
> Are you saying the core already sets the STP to disabled, for ports
> which are unused? I did not spot that in your previous patch?
Just look at dsa_port_disable Andrew:
void dsa_port_disable(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
if (!dp->bridge_dev)
dsa_port_set_state_now(dp, BR_STATE_DISABLED);
if (ds->ops->port_disable)
ds->ops->port_disable(ds, port);
}
The only thing worth arguing here is whether it makes sense to call
ds->ops->disable for a bridged port, or should we simply return right
away in this case. But this would be an independent patch anyway.
Thank you,
Vivien
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox