kernel.git/drivers/net/ipa/ipa_endpoint.c, branch linux-5.8.y

net: ipa: skip suspend/resume activities if not set up

2020-10-29T09:07:00Z

[ Upstream commit d1704382821032fede445b816f4296fd379baacf ] When processing a system suspend request we suspend modem endpoints if they are enabled, and call ipa_cmd_tag_process() (which issues IPA commands) to ensure the IPA pipeline is cleared. It is an error to attempt to issue an IPA command before setup is complete, so this is clearly a bug. But we also shouldn't suspend or resume any endpoints that have not been set up. Have ipa_endpoint_suspend() and ipa_endpoint_resume() immediately return if setup hasn't completed, to avoid any attempt to configure endpoints or issue IPA commands in that case. Fixes: 84f9bd12d46d ("soc: qcom: ipa: IPA endpoints") Tested-by: Matthias Kaehlcke Signed-off-by: Alex Elder Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

net: ipa: introduce ipa_cmd_tag_process()

2020-06-30T20:10:57Z

Create a new function ipa_cmd_tag_process() that simply allocates a transaction, adds a tag process command to it to clear the hardware pipeline, and commits the transaction. Call it in from ipa_endpoint_suspend(), after suspending the modem endpoints but before suspending the AP command TX and AP LAN RX endpoints (which are used by the tag sequence). Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: header pad field only valid for AP->modem endpoint

2020-06-12T01:39:08Z

Only QMAP endpoints should be configured to find a pad size field within packet headers. They are found in the first byte of the QMAP header (and the hardware fills only the 6 bits in that byte that constitute the pad_len field). The RMNet driver assumes the pad_len field is valid for received packets, so we want to ensure the pad_len field is filled in that case. That driver also assumes the length in the QMAP header includes the pad bytes. The RMNet driver does *not* pad the packets it sends, so the pad_len field can be ignored. Fix ipa_endpoint_init_hdr_ext() so it only marks the pad field offset valid for QMAP RX endpoints, and in that case indicates that the length field in the header includes the pad bytes. Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: program upper nibbles of sequencer type

2020-06-12T01:39:08Z

The upper two nibbles of the sequencer type were not used for SDM845, and were assumed to be 0. But for SC7180 they are used, and so they must be programmed by ipa_endpoint_init_seq(). Fix this bug. IPA_SEQ_PKT_PROCESS_NO_DEC_NO_UCP_DMAP doesn't have a descriptive comment, so add one. Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: program metadata mask differently

2020-06-12T01:39:07Z

The way the mask value is programmed for QMAP RX endpoints was based on some wrong assumptions about the way metadata containing the QMAP mux_id value is formatted. The metadata value supplied by the modem is *not* in QMAP format, and in fact contains the mux_id we want in its (big endian) low-order byte. That byte must be written by the IPA into offset 1 of the QMAP header it inserts before the received packet. QMAP TX endpoints *do* use a QMAP header as the metadata sent with each packet. The modem assumes this, and based on that assumes the mux_id is in the second byte. To match those assumptions we must program the modem TX (QMAP) endpoint HDR register to indicate the metadata will be found at offset 0 in the message header. The previous configuration managed to work, but it was not working correctly. This patch fixes a bug whose symptom was receipt of messages containing the wrong QMAP mux_id. In fixing this, get rid of ipa_rmnet_mux_id_metadata_mask(), which was more or less defined so there was a separate place to explain what was happening as we generated the mask value. Instead, put a longer description of how this works above ipa_endpoint_init_hdr(), and define the metadata mask to use as a simple constant. Signed-off-by: Alex Elder Signed-off-by: David S. Miller

drivers: ipa: fix typoes for ipa

2020-05-27T21:58:28Z

Change "transactio" -> "transaction". Also an alignment correction. Signed-off-by: Wang Wenhu Signed-off-by: David S. Miller

net: ipa: Remove ipa_endpoint_stop{,_rx_dma} again

2020-05-09T05:35:56Z

When building arm64 allyesconfig: drivers/net/ipa/ipa_endpoint.c: In function 'ipa_endpoint_stop_rx_dma': drivers/net/ipa/ipa_endpoint.c:1274:13: error: 'IPA_ENDPOINT_STOP_RX_SIZE' undeclared (first use in this function) drivers/net/ipa/ipa_endpoint.c:1274:13: note: each undeclared identifier is reported only once for each function it appears in drivers/net/ipa/ipa_endpoint.c:1289:2: error: implicit declaration of function 'ipa_cmd_dma_task_32b_addr_add' [-Werror=implicit-function-declaration] drivers/net/ipa/ipa_endpoint.c:1291:45: error: 'ENDPOINT_STOP_DMA_TIMEOUT' undeclared (first use in this function) drivers/net/ipa/ipa_endpoint.c: In function 'ipa_endpoint_stop': drivers/net/ipa/ipa_endpoint.c:1309:16: error: 'IPA_ENDPOINT_STOP_RX_RETRIES' undeclared (first use in this function) These functions were removed in a series, merged in as commit 33395f4a5c1b ("Merge branch 'net-ipa-kill-endpoint-stop-workaround'"). Remove them again so that the build works properly. Fixes: 3793faad7b5b ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net") Signed-off-by: Nathan Chancellor Reviewed-by: Arnd Bergmann Signed-off-by: Jakub Kicinski

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2020-05-07T05:10:13Z

Conflicts were all overlapping changes. Signed-off-by: David S. Miller

net: ipa: kill ipa_endpoint_stop()

2020-05-07T00:41:32Z

The previous commit made ipa_endpoint_stop() be a trivial wrapper around gsi_channel_stop(). Since it no longer does anything special, just open-code it in the three places it's used. Signed-off-by: Alex Elder Signed-off-by: David S. Miller

net: ipa: don't retry in ipa_endpoint_stop()

2020-05-07T00:41:11Z

The only reason ipa_endpoint_stop() had a retry loop was that the just-removed workaround required an IPA DMA command to occur between attempts. The gsi_channel_stop() call that implements the stop does its own retry loop, to cover a channel's transition from started to stop-in-progress to stopped state. Get rid of the unnecessary retry loop in ipa_endpoint_stop(). Signed-off-by: Alex Elder Signed-off-by: David S. Miller