linux-toradex.git/drivers/net/vmxnet3, branch v6.5

net: introduce and use skb_frag_fill_page_desc()

2023-05-13T18:47:56+00:00

Most users use __skb_frag_set_page()/skb_frag_off_set()/
skb_frag_size_set() to fill the page desc for a skb frag.

Introduce skb_frag_fill_page_desc() to do that.

net/bpf/test_run.c does not call skb_frag_off_set() to
set the offset, "copy_from_user(page_address(page), ...)"
and 'shinfo' being part of the 'data' kzalloced in
bpf_test_init() suggest that it is assuming offset to be
initialized as zero, so call skb_frag_fill_page_desc()
with offset being zero for this case.

Also, skb_frag_set_page() is not used anymore, so remove
it.

Signed-off-by: Yunsheng Lin 
Reviewed-by: Leon Romanovsky 
Reviewed-by: Simon Horman 
Signed-off-by: David S. Miller

net: vmxnet3: Fix NULL pointer dereference in vmxnet3_rq_rx_complete()

2023-04-19T08:03:05+00:00

When vmxnet3_rq_create() fails to allocate rq->data_ring.base due to page
allocation failure, subsequent call to vmxnet3_rq_rx_complete() can result in
NULL pointer dereference.

To fix this bug, check not only that rxDataRingUsed is true but also that
adapter->rxdataring_enabled is true before calling memcpy() in
vmxnet3_rq_rx_complete().

[1728352.477993] ethtool: page allocation failure: order:9, mode:0x6000c0(GFP_KERNEL), nodemask=(null),cpuset=/,mems_allowed=0
...
[1728352.478009] Call Trace:
[1728352.478028]  dump_stack+0x41/0x60
[1728352.478035]  warn_alloc.cold.120+0x7b/0x11b
[1728352.478038]  ? _cond_resched+0x15/0x30
[1728352.478042]  ? __alloc_pages_direct_compact+0x15f/0x170
[1728352.478043]  __alloc_pages_slowpath+0xcd3/0xd10
[1728352.478047]  __alloc_pages_nodemask+0x2e2/0x320
[1728352.478049]  __dma_direct_alloc_pages.constprop.25+0x8a/0x120
[1728352.478053]  dma_direct_alloc+0x5a/0x2a0
[1728352.478056]  vmxnet3_rq_create.part.57+0x17c/0x1f0 [vmxnet3]
...
[1728352.478188] vmxnet3 0000:0b:00.0 ens192: rx data ring will be disabled
...
[1728352.515347] BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
...
[1728352.515440] RIP: 0010:memcpy_orig+0x54/0x130
...
[1728352.515655] Call Trace:
[1728352.515665]  
[1728352.515672]  vmxnet3_rq_rx_complete+0x419/0xef0 [vmxnet3]
[1728352.515690]  vmxnet3_poll_rx_only+0x31/0xa0 [vmxnet3]
...

Signed-off-by: Seiji Nishikawa 
Signed-off-by: David S. Miller

vmxnet3: use gro callback when UPT is enabled

2023-03-25T02:13:49+00:00

Currently, vmxnet3 uses GRO callback only if LRO is disabled. However,
on smartNic based setups where UPT is supported, LRO can be enabled
from guest VM but UPT devicve does not support LRO as of now. In such
cases, there can be performance degradation as GRO is not being done.

This patch fixes this issue by calling GRO API when UPT is enabled. We
use updateRxProd to determine if UPT mode is active or not.

To clarify few things discussed over the thread:
The patch is not neglecting any feature bits nor disabling GRO. It uses
GRO callback when UPT is active as LRO is not available in UPT.
GRO callback cannot be used as default for all cases as it degrades
performance for non-UPT cases or for cases when LRO is already done in
ESXi.

Cc: stable@vger.kernel.org
Fixes: 6f91f4ba046e ("vmxnet3: add support for capability registers")
Signed-off-by: Ronak Doshi 
Reviewed-by: Simon Horman 
Link: https://lore.kernel.org/r/20230323200721.27622-1-doshir@vmware.com
Signed-off-by: Jakub Kicinski

vmxnet3: move rss code block under eop descriptor

2023-02-10T06:50:46+00:00

Commit b3973bb40041 ("vmxnet3: set correct hash type based on
rss information") added hashType information into skb. However,
rssType field is populated for eop descriptor. This can lead
to incorrectly reporting of hashType for packets which use
multiple rx descriptors. Multiple rx descriptors are used
for Jumbo frame or LRO packets, which can hit this issue.

This patch moves the RSS codeblock under eop descritor.

Cc: stable@vger.kernel.org
Fixes: b3973bb40041 ("vmxnet3: set correct hash type based on rss information")
Signed-off-by: Ronak Doshi 
Acked-by: Peng Li 
Acked-by: Guolin Yang 
Link: https://lore.kernel.org/r/20230208223900.5794-1-doshir@vmware.com
Signed-off-by: Jakub Kicinski

vmxnet3: correctly report csum_level for encapsulated packet

2022-12-22T01:55:30+00:00

Commit dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload
support") added support for encapsulation offload. However, the
pathc did not report correctly the csum_level for encapsulated packet.

This patch fixes this issue by reporting correct csum level for the
encapsulated packet.

Fixes: dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload support")
Signed-off-by: Ronak Doshi 
Acked-by: Peng Li 
Link: https://lore.kernel.org/r/20221220202556.24421-1-doshir@vmware.com
Signed-off-by: Jakub Kicinski

vmxnet3: use correct intrConf reference when using extended queues

2022-12-02T10:30:07+00:00

Commit 39f9895a00f4 ("vmxnet3: add support for 32 Tx/Rx queues")
added support for 32Tx/Rx queues. As a part of this patch, intrConf
structure was extended to incorporate increased queues.

This patch fixes the issue where incorrect reference is being used.

Fixes: 39f9895a00f4 ("vmxnet3: add support for 32 Tx/Rx queues")
Signed-off-by: Ronak Doshi 
Acked-by: Guolin Yang 
Signed-off-by: David S. Miller

vmxnet3: correctly report encapsulated LRO packet

2022-12-02T10:30:07+00:00

Commit dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload
support") added support for encapsulation offload. However, the
pathc did not report correctly the encapsulated packet which is
LRO'ed by the hypervisor.

This patch fixes this issue by using correct callback for the LRO'ed
encapsulated packet.

Fixes: dacce2be3312 ("vmxnet3: add geneve and vxlan tunnel offload support")
Signed-off-by: Ronak Doshi 
Acked-by: Guolin Yang 
Signed-off-by: David S. Miller

net: drop the weight argument from netif_napi_add

2022-09-29T01:57:14+00:00

We tell driver developers to always pass NAPI_POLL_WEIGHT
as the weight to netif_napi_add(). This may be confusing
to newcomers, drop the weight argument, those who really
need to tweak the weight can use netif_napi_add_weight().

Acked-by: Marc Kleine-Budde  # for CAN
Link: https://lore.kernel.org/r/20220927132753.750069-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski

net: move from strlcpy with unused retval to strscpy

2022-08-31T21:11:07+00:00

Follow the advice of the below link and prefer 'strscpy' in this
subsystem. Conversion is 1:1 because the return value is not used.
Generated by a coccinelle script.

Link: https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/
Signed-off-by: Wolfram Sang 
Acked-by: Marc Kleine-Budde  # for CAN
Link: https://lore.kernel.org/r/20220830201457.7984-1-wsa+renesas@sang-engineering.com
Signed-off-by: Jakub Kicinski

vmxnet3: do not reschedule napi for rx processing

2022-07-29T11:11:23+00:00

Commit '2c5a5748105a ("vmxnet3: add support for out of order rx
completion")' added support for out of order rx completion. Within
that patch, an enhancement was done to reschedule napi for processing
rx completions.

However, it can lead to missing an interrupt. So, this patch reverts
that part of the code.

Fixes: 2c5a5748105a ("vmxnet3: add support for out of order rx completion")
Signed-off-by: Ronak Doshi 
Acked-by: Guolin Yang 
Signed-off-by: David S. Miller