kernel.git/drivers/soc/qcom/mdt_loader.c, branch linux-6.4.y

soc: qcom: mdt_loader: Fix unconditional call to scm_pas_mem_setup

2023-07-23T11:53:57Z

commit bcb889891371c3cf767f2b9e8768cfe2fdd3810f upstream. Commit ebeb20a9cd3f ("soc: qcom: mdt_loader: Always invoke PAS mem_setup") dropped the relocate check and made pas_mem_setup run unconditionally. The code was later moved with commit f4e526ff7e38 ("soc: qcom: mdt_loader: Extract PAS operations") to qcom_mdt_pas_init() effectively losing track of what was actually done. The assumption that PAS mem_setup can be done anytime was effectively wrong, with no good reason and this caused regression on some SoC that use remoteproc to bringup ath11k. One example is IPQ8074 SoC that effectively broke resulting in remoteproc silently die and ath11k not working. On this SoC FW relocate is not enabled and PAS mem_setup was correctly skipped in previous kernel version resulting in correct bringup and function of remoteproc and ath11k. To fix the regression, reintroduce the relocate check in qcom_mdt_pas_init() and correctly skip PAS mem_setup where relocate is not enabled. Fixes: ebeb20a9cd3f ("soc: qcom: mdt_loader: Always invoke PAS mem_setup") Tested-by: Robert Marko Co-developed-by: Robert Marko Signed-off-by: Robert Marko Signed-off-by: Christian Marangi Cc: stable@vger.kernel.org Reviewed-by: Mukesh Ojha Signed-off-by: Bjorn Andersson Link: https://lore.kernel.org/r/20230526115511.3328-1-ansuelsmth@gmail.com Signed-off-by: Greg Kroah-Hartman

firmware: qcom_scm: Move qcom_scm.h to include/linux/firmware/qcom/

2023-02-09T03:15:16Z

Move include/linux/qcom_scm.h to include/linux/firmware/qcom/qcom_scm.h. This removes 1 of a few remaining Qualcomm-specific headers into a more approciate subdirectory under include/. Suggested-by: Bjorn Andersson Signed-off-by: Elliot Berman Reviewed-by: Guru Das Srinagesh Acked-by: Mukesh Ojha Signed-off-by: Bjorn Andersson Link: https://lore.kernel.org/r/20230203210956.3580811-1-quic_eberman@quicinc.com

soc: qcom: correct kerneldoc

2022-06-26T03:05:31Z

Correct kerneldoc warnings like: drivers/soc/qcom/mdt_loader.c:126: warning: Function parameter or member 'fw_name' not described in 'qcom_mdt_read_metadata' Signed-off-by: Krzysztof Kozlowski Reviewed-by: Stephen Boyd Signed-off-by: Bjorn Andersson Link: https://lore.kernel.org/r/20220519073301.7072-2-krzysztof.kozlowski@linaro.org

soc: qcom: mdt_loader: Fix split-firmware condition

2022-02-23T19:13:12Z

The updated condition checking if a segment can be found in the loaded firmware blob, or need to be loaded from a separate file, incorrectly classifies segments that ends at the end of the loaded blob. The result is that the mdt loader attempts to load the segment from a separate file. Correct the conditional to use the loaded segment instead. Fixes: ea90330fa329 ("soc: qcom: mdt_loader: Extend check for split firmware") Signed-off-by: Bjorn Andersson Reviewed-by: Vinod Koul Link: https://lore.kernel.org/r/20220215034819.1209367-1-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Extract PAS operations

2022-02-04T03:54:48Z

Rather than passing a boolean to indicate if the PAS operations should be performed from within __mdt_load(), extract them to their own helper function. This will allow clients to invoke this directly, with some qcom_scm_pas_metadata context that they later needs to release, without further having to complicate the prototype of qcom_mdt_load(). Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-9-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Always invoke PAS mem_setup

2022-02-03T21:32:30Z

After spelunking various old kernel trees no finds has been found indicating that the PAS mem_setup call should actually be made conditional on the image being relocatable. Group the two PAS operations together, to facilitate splitting them out in a following patch. Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-8-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Reorder parts of __qcom_mdt_load()

2022-02-03T21:32:30Z

Move the traversal of the program headers to the start of the function, to make sure that min_ and max_addr are in scope as the call to qcom_scm_pas_mem_setup() is moved in the next commit. Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-7-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Extend check for split firmware

2022-02-03T21:32:30Z

Some of the Qualcomm SM8450 firmware files are padded such that the start of the first segment falls within the .mdt file but the segment to be loaded is stored as a separate .bNN file. Extend the condition to only attempt to read a segment inline if the entire segment would be available. Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-6-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Allow hash to reside in any segment

2022-02-03T21:32:30Z

It's been observed that some firmware found on Qualcomm SM8450 devices carries the hash segment as the last segment in the ELF. Extend the support to allow picking the hash from any segment in the MDT/MBN. Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-5-bjorn.andersson@linaro.org

soc: qcom: mdt_loader: Allow hash segment to be split out

2022-02-03T21:32:30Z

It's been observed that some firmware found in a Qualcomm SM8450 device has the hash table in a separate .bNN file. Use the newly extracted helper function to load this segment from the separate file, if it's determined that the hashes are not part of the already loaded firmware. In order to do this, the function needs access to the firmware basename and to provide more useful error messages a struct device to associate the errors with. Signed-off-by: Bjorn Andersson Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20220128025513.97188-4-bjorn.andersson@linaro.org