sc846: add remoteproc support - #3499
Draft
artursartamonovsadi wants to merge 3 commits into
Draft
artursartamonovsadi wants to merge 3 commits into
artursartamonovsadi wants to merge 3 commits into
Conversation
Update the SC846 SoM SHARC-FX node for the new remoteproc interface, including reset, system-vector, mailbox and vring configuration. Correct the L1 and L2 address ranges and relocate the resource table and vring into L2 memory. Signed-off-by: Arturs Artamonovs <Arturs.Artamonovs@analog.com>
Allow ADI_REMOTEPROC to be selected on ARCH_SC84X platforms so their SHARC DSP cores can use the ADI remoteproc driver. Signed-off-by: Arturs Artamonovs <Arturs.Artamonovs@analog.com>
Enable the ADI remoteproc driver to load ELF firmware alongside existing LDR images. Translate SHARC device addresses to physical memory and use MDMA to load segments and clear zero-filled regions, avoiding unsupported narrow ARM accesses to SHARC-FX instruction RAM. Set the system vector from the ELF entry point and use an embedded resource table when available. Fall back to the driver's built-in table for firmware without a resource table so RPMsg remains supported. Also correct L1 device-to-virtual address translation. Signed-off-by: Arturs Artamonovs <Arturs.Artamonovs@analog.com>
ukleinek
reviewed
Sep 9, 2026
ukleinek
left a comment
ukleinek
left a comment
Collaborator
There was a problem hiding this comment.
I didn't test this yet on sc5xx (but intent to). Just some general, mostly nitpick comments.
| u32 verify; | ||
| struct adi_sharc_resource_table *adi_rsc_table; | ||
| struct sharc_resource_table *loaded_rsc_table; | ||
| /* True when the resource table came from the firmware image itself |
Collaborator
There was a problem hiding this comment.
Suggested change
| /* True when the resource table came from the firmware image itself | |
| /* | |
| * True when the resource table came from the firmware image itself |
Comment on lines
+550
to
+559
| * adi_elf_load_segments: load ELF PT_LOAD segments over MDMA | ||
| * | ||
| * Mirrors rproc_elf_load_segments() but routes every write through MDMA | ||
| * rather than memcpy(), because IRAM cannot take narrow accesses from the | ||
| * ARM. See adi_rproc_dma_write(). | ||
| * | ||
| * The SPU is held open for the duration of the load, as the LDR path does: | ||
| * the SHARC-FX bus completer ports reject non-secure accesses with an error | ||
| * response rather than completing them. | ||
| */ |
Collaborator
There was a problem hiding this comment.
I don't see how, but this would be considerably easier to mainline if it could reuse rproc_elf_load_segments() somehow (or a generic extension of it).
Author
There was a problem hiding this comment.
I will try to look into it. As far as I understood problem is that Linux is non-secure level, and L2 access to SHARC-FX is only allowed to use with MDMA from non-secure state, when its also set in SPU to accept non secure access. MDMA SPU configured from U-boot
| * ELF has no equivalent pass, and without this ldr_load_addr would | ||
| * still hold SHARC_IDLE_ADDR from probe, leaving the core spinning in | ||
| * the bootrom idle loop instead of running the loaded firmware. Use | ||
| * the entry point remoteproc resolved through .get_boot_addr. |
Collaborator
There was a problem hiding this comment.
Suggested change
| * the entry point remoteproc resolved through .get_boot_addr. | |
| * the entry point remoteproc resolved through .get_boot_addr(). |
Comment on lines
-555
to
+893
| if (ret < 0) | ||
| if (ret < 0) { | ||
| return ret; | ||
| } |
Collaborator
There was a problem hiding this comment.
This hunk should be dropped.
Comment on lines
-781
to
+1124
| if (da >= rproc_data->l1_da_range[0] && da < rproc_data->l1_da_range[1]) | ||
| ret = L1_shared_base + da; | ||
| else if (da >= rproc_data->l2_da_range[0] && da < rproc_data->l2_da_range[1]) | ||
| if (da >= rproc_data->l1_da_range[0] && da < rproc_data->l1_da_range[1]) { | ||
| ret = L1_shared_base + (da - rproc_data->l1_da_range[0]); | ||
| } | ||
| else if (da >= rproc_data->l2_da_range[0] && da < rproc_data->l2_da_range[1]) { | ||
| ret = L2_shared_base + (da - rproc_data->l2_da_range[0]); | ||
| } |
Collaborator
There was a problem hiding this comment.
This only adds curly braces that are not welcome in the kernel coding style. Please drop this hunk.
| &svect_args); | ||
| if (ret) | ||
| return dev_err_probe(dev, ret, "Missing adi,svect property\n"); | ||
|
|
Collaborator
There was a problem hiding this comment.
I would keep this empty line.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR Description
Add support for remoteproc support on sc846 platform
PR Type
PR Checklist