Skip to content

zephyr: userspace_helper: add test_mailbox.c (plus a test) #10477

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kv2019i merged 2 commits into from
Jan 23, 2026
Merged

zephyr: userspace_helper: add test_mailbox.c (plus a test) #10477

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8d7ceea
zephyr: userspace_helper: add user_access_to_mailbox()
kv2019i Jan 9, 2026
bf8535a
zephyr: test: userspace: add test_mailbox for sof/mailbox.h
kv2019i Jan 9, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions zephyr/include/rtos/userspace_helper.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,4 +91,23 @@ int user_stack_free(void *p_stack);
*/
void module_driver_heap_remove(struct k_heap *mod_drv_heap);

#ifdef CONFIG_USERSPACE

/**
* Add access to mailbox.h interface to a user-space thread.
*
* @param domain memory domain to add the mailbox partitions to
* @param thread_id user-space thread for which access is added
*/
int user_access_to_mailbox(struct k_mem_domain *domain, k_tid_t thread_id);

#else

static inline int user_access_to_mailbox(struct k_mem_domain *domain, k_tid_t thread_id)
{
return 0;
}

#endif /* CONFIG_USERSPACE */

#endif /* __ZEPHYR_LIB_USERSPACE_HELPER_H__ */
41 changes: 41 additions & 0 deletions zephyr/lib/userspace_helper.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@
#include <rtos/userspace_helper.h>
#include <sof/audio/module_adapter/module/generic.h>
#include <sof/audio/module_adapter/library/userspace_proxy.h>
#include <sof/lib/mailbox.h>

#define MODULE_DRIVER_HEAP_CACHED CONFIG_SOF_ZEPHYR_HEAP_CACHED

Expand Down Expand Up @@ -82,6 +83,46 @@ int user_memory_init_shared(k_tid_t thread_id, struct processing_module *mod)
return k_mem_domain_add_thread(comp_dom, thread_id);
}

int user_access_to_mailbox(struct k_mem_domain *domain, k_tid_t thread_id)
{
struct k_mem_partition mem_partition;
int ret;

/*
* Start with mailbox_swregs. This is aligned with mailbox.h
* implementation with uncached addressed used for register I/O.
*/
mem_partition.start =
(uintptr_t)sys_cache_uncached_ptr_get((void __sparse_cache *)MAILBOX_SW_REG_BASE);

BUILD_ASSERT(MAILBOX_SW_REG_SIZE == CONFIG_MMU_PAGE_SIZE);
mem_partition.size = CONFIG_MMU_PAGE_SIZE;
mem_partition.attr = K_MEM_PARTITION_P_RW_U_RW;

ret = k_mem_domain_add_partition(domain, &mem_partition);
Copy link
Copy Markdown
Collaborator

@lyakh lyakh Jan 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kv2019i oh, wait... don't we need a method to remove the partition too? This is for the LL scheduling thread, right? Would it be restarted during the life time of the firmware? If so, then we need to remove the partition too.

if (ret < 0)
return ret;

#ifndef CONFIG_IPC_MAJOR_4
Copy link
Copy Markdown
Collaborator

@lyakh lyakh Jan 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Userspace with IPC3? really?

Copy link
Copy Markdown
Collaborator Author

@kv2019i kv2019i Jan 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@lyakh Probably no, but there is no technical relation with userspace and IPC protocol version. I did this test with a brute force check of how mailbox.h is used in SOF application code and tried to cover all usages in this test PR. If I assume some functions are only used with IPC4 (see #10480 ), and then proceed to to only test with another assumption that user-space is only used in IPC4, I'm not sure if future SOF maintainers will appreciate this. OTOH, I can just drop this as well -- this won't prevent IPC3 usage either.

Copy link
Copy Markdown
Member

@abonislawski abonislawski Jan 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cool, long life support for IPC3 :)

Copy link
Copy Markdown
Collaborator

@lyakh lyakh Jan 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kv2019i My concern is purely practical - dead code. Has anyone ever tested userspace with IPC3? I suspect (and hope) that not. And I estimate the probability of this ever being used as very low, but you might know something that I don't! So, if I'm right and there's no somewhat realistic chance of this being used - please let's remove this.

/*
* Next mailbox_stream (not available in IPC4). Stream access is cached,
* so different mapping this time.
*/
mem_partition.start =
(uintptr_t)sys_cache_cached_ptr_get((void *)SRAM_STREAM_BASE);
BUILD_ASSERT(MAILBOX_STREAM_SIZE == CONFIG_MMU_PAGE_SIZE);
/* size and attr the same as for mailbox_swregs */

ret = k_mem_domain_add_partition(domain, &mem_partition);
if (ret < 0)
return ret;
#endif

k_mem_domain_add_thread(domain, thread_id);

return 0;
}

#else /* CONFIG_USERSPACE */

void *user_stack_allocate(size_t stack_size, uint32_t options)
Expand Down