Hosts the 0x221E linux distro kernel. https://universe.0xinfinity.dev/distro/kernel/atom?h=linux-5.1.y 2019-02-07T17:52:39Z 2019-02-07T17:52:39Z Thierry Reding treding@nvidia.com 2019-02-07T11:50:06Z urn:sha1:fe45ab552955ee93b492cbf86603eb02b3c77a66 The BPMP implementation on Tegra194 is mostly compatible with the implementation on Tegra186, so make sure the latter is available when support for Tegra194 is enabled. Suggested-by: Timo Alho <talho@nvidia.com> Reviewed-by: Timo Alho <talho@nvidia.com> Tested-by: Timo Alho <talho@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2019-02-07T17:52:19Z Thierry Reding treding@nvidia.com 2019-02-07T11:50:05Z urn:sha1:79d031fcad56e27fc4d614d54cd5962cad282473 Only include support for Tegra210 and Tegra186 in the BPMP driver if support for those SoCs was selected. This fixes a build failure seen on 32-bit ARM allmodconfig builds, but could also happen on 64-bit ARM builds if either Tegra210 or Tegra186 were not selected. Reported-by: Guenter Roeck <linux@roeck-us.net> Reviewed-by: Timo Alho <talho@nvidia.com> Tested-by: Timo Alho <talho@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2019-01-25T14:58:47Z Thierry Reding treding@nvidia.com 2019-01-25T14:46:56Z urn:sha1:e835d442cff641a0831b090a47457366d7919324 Many of the include files are not needed, so drop them. Signed-off-by: Thierry Reding <treding@nvidia.com> 2019-01-25T14:58:47Z Timo Alho talho@nvidia.com 2019-01-24T17:03:54Z urn:sha1:139251fc220830cc49b71331d281a8ad03a08ab7 This patch adds driver for Tegra210 BPMP firmware. The BPMP is a specific processor in Tegra210 chip, which runs firmware for assisting in entering deep low power states (suspend to ram), and offloading DRAM memory clock scaling on some platforms. Based on work by Sivaram Nair <sivaramn@nvidia.com> Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2019-01-25T14:58:47Z Timo Alho talho@nvidia.com 2019-01-24T17:03:53Z urn:sha1:cdfa358b248efd36c6a9cb4d4d0a3ba7509f8387 Split BPMP driver into common and chip specific parts to facilitate adding support for previous and future Tegra chips that are using BPMP as co-processor. Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2019-01-25T14:57:52Z Timo Alho talho@nvidia.com 2019-01-24T17:03:52Z urn:sha1:165ce6e01d2d00bff2f10eba8dd679cee44163bd As a preparatory change to refactor BPMP driver to support other than Tegra186 and Tegra194 chip generations, reword and slightly refactor some of the functions to better match with what is actually happening in the wire-level protocol. The communication with BPMP is essentially a Remote Procedure Call consisting of "request" and "response". Either side (BPMP or CPU) can initiate the communication. The state machine for communication consists of following steps (from Linux point of view): Linux initiating the call: 1) check that channel is free to transmit a request (is_request_channel_free) 2) copy request message payload to shared location 3) post the request in channel (post_request) 4) notify BPMP that channel state has been updated (ring_doorbell) 5) wait for response (is_response_ready) 6) copy response message payload from shared location 7) acknowledge the response in channel (ack_response) BPMP initiating the call: 1) wait for request (is_request_ready) 2) copy request message payload from shared location 3) acknowledge the request in channel (ack_request) 4) check that channel is free to transmit response (is_response_channel_free) 5) copy response message payload to shared location 6) post the response message to channel (post_response) 7) notify BPMP that channel state has been updated (ring_doorbell) Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2018-11-08T11:49:26Z Timo Alho talho@nvidia.com 2018-10-22T13:19:40Z urn:sha1:af51c25fb59d4365c259ab93f388462e0998a2ed Add support for a new MRQ, that uses in-band messaging instead of IOVA buffer, to retrieve the firmware version 'tag' during boot. If an older firmware is used, that does not support the new MRQ, fall back to the earlier implementation. Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Sivaram Nair <sivaramn@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2018-11-08T11:49:26Z Timo Alho talho@nvidia.com 2018-10-22T13:19:38Z urn:sha1:2b86c11b99d3fc4e82af65c5b00b4334a6dfe6c9 Last two characters of the version tag that is 32 bytes long were stripped out. Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Sivaram Nair <sivaramn@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2018-11-08T11:49:25Z Timo Alho talho@nvidia.com 2018-10-22T13:19:37Z urn:sha1:43dc7485807a8ffde12e242e557a1992478de4de Patch "firmware: tegra: add helper to check for supported MRQs" added an API to check if MRQ is supported. Remove the implementation from bpmp debugfs code in favor of that. Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Sivaram Nair <sivaramn@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com> 2018-11-08T11:49:25Z Timo Alho talho@nvidia.com 2018-10-22T13:19:36Z urn:sha1:d78b5bde0ffc33d20f014b3ad4d7aaac8b79d34e Add a helper function to check that firmware is supporting a given MRQ command. Signed-off-by: Timo Alho <talho@nvidia.com> Acked-by: Sivaram Nair <sivaramn@nvidia.com> Acked-by: Jon Hunter <jonathanh@nvidia.com> Signed-off-by: Thierry Reding <treding@nvidia.com>