kernel/linux.git/drivers/net/ipa, branch v5.18.2

net: ipa: compute proper aggregation limit

2022-06-06T06:48:54+00:00

commit c5794097b269f15961ed78f7f27b50e51766dec9 upstream. The aggregation byte limit for an endpoint is currently computed based on the endpoint's receive buffer size. However, some bytes at the front of each receive buffer are reserved on the assumption that--as with SKBs--it might be useful to insert data (such as headers) before what lands in the buffer. The aggregation byte limit currently doesn't take into account that reserved space, and as a result, aggregation could require space past that which is available in the buffer. Fix this by reducing the size used to compute the aggregation byte limit by the NET_SKB_PAD offset reserved for each receive buffer. Signed-off-by: Alex Elder Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

net: ipa: get rid of a duplicate initialization

2022-05-13T11:01:42+00:00

In ipa_qmi_ready(), the "ipa" local variable is set when initialized, but then set again just before it's first used. One or the other is enough, so get rid of the first one. References: https://lore.kernel.org/lkml/200de1bd-0f01-c334-ca18-43eed783dfac@intel.com/ Reported-by: kernel test robot Fixes: 530f9216a953 ("soc: qcom: ipa: AP/modem communications") Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: record proper RX transaction count

2022-05-13T11:01:42+00:00

Each time we are notified that some number of transactions on an RX channel has completed, we record the number of bytes that have been transferred since the previous notification. We also track the number of transactions completed, but that is not currently being calculated correctly; we're currently counting the number of such notifications, but each notification can represent many transaction completions. Fix this. Fixes: 650d1603825d8 ("soc: qcom: ipa: the generic software interface") Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: certain dropped packets aren't accounted for

2022-05-13T11:01:42+00:00

If an RX endpoint receives packets containing status headers, and a packet in the buffer is not dropped, ipa_endpoint_skb_copy() is responsible for wrapping the packet data in an SKB and forwarding it to ipa_modem_skb_rx() for further processing. If ipa_endpoint_skb_copy() gets a null pointer from build_skb(), it just returns early. But in the process it doesn't record that as a dropped packet in the network device statistics. Instead, call ipa_modem_skb_rx() whether or not the SKB pointer is NULL; that function ensures the statistics are properly updated. Fixes: 1b65bbcc9a710 ("net: ipa: skip SKB copy if no netdev") Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: use struct_size() for the interconnect array

2022-03-12T06:50:07+00:00

In review for commit 8ee7ec4890e2b ("net: ipa: embed interconnect array in the power structure"), Jakub Kicinski suggested that a follow-up patch use struct_size() when computing the size of the IPA power structure, which ends with a flexible array member. Do that. Suggested-by: Jakub Kicinski Signed-off-by: Alex Elder Link: https://lore.kernel.org/r/20220311162423.872645-1-elder@linaro.org Signed-off-by: Jakub Kicinski

net: ipa: use IPA power device pointer

2022-03-11T05:19:07+00:00

The ipa_power structure contains a copy of the IPA device pointer, so there's no need to pass it to ipa_interconnect_init(). We can also use that pointer for an error message in ipa_power_enable(). Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski

net: ipa: embed interconnect array in the power structure

2022-03-11T05:19:06+00:00

Rather than allocating the interconnect array dynamically, represent the interconnects with a variable-length array at the end of the ipa_power structure. Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski

net: ipa: use bulk interconnect initialization

2022-03-11T05:19:06+00:00

The previous patch used bulk interconnect operations to initialize IPA interconnects one at a time. This rearranges things to use the bulk interfaces as intended--on all interconnects together. As a result ipa_interconnect_init_one() and ipa_interconnect_exit_one() are no longer needed. Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski

net: ipa: use bulk operations to set up interconnects

2022-03-11T05:19:05+00:00

Use of_icc_bulk_get() and icc_bulk_put(), icc_bulk_set_bw(), and icc_bulk_enable() and icc_bulk_disable() to initialize individual IPA interconnects. Those functions already log messages in the event of error so we don't need to. Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski

net: ipa: use interconnect bulk enable/disable operations

2022-03-11T05:19:05+00:00

The power interconnect array is now an array of icc_bulk_data structures, which is what the interconnect bulk enable and disable functions require. Get rid of ipa_interconnect_enable() and ipa_interconnect_disable(), and just call icc_bulk_enable() and icc_bulk_disable() instead. Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski