core: interrupt: move to new interrupt API functions and remove the old ones

[etienne-lms]

10 patches extracted from #5954 + a fixup patch (for series 1st commit) to finalise new interrupt management API functions that can be used on CPU main interrupt controller as well as on secondary interrupt controllers.

This series makes main controller to use generic resources, then it moves each interrupt consumer to the new API functions, last the series removes old interrupt API functions itr_*().

1st commit (("drivers: move to interrupt_call_handlers()") has a review tag applied but I appended a fixup commit for some rewording in inline description comments.

[jenswikl]

If you don't mind, please wait with "[review] drivers: move to interrupt_call_handlers" until we see how the bottom half handlers will play out.

[etienne-lms]

Ok, i'll drop the fixup commit.

[etienne-lms]

@borneoa, you may want to have a look at this series.

[etienne-lms]

Updated with interrupt_raise_pi() and friends API functions defined in interrupt.h.

[jenswikl]

For commit "core: interrupt: helper function for raise_pi, raise_sgi, set_affinity", please apply:
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

[etienne-lms]

May I squash the fixup commit and reorder the patches?

[jenswikl]

May I squash the fixup commit and reorder the patches?

Yes, please.

[etienne-lms]

Fixup commits squashed and commits reordered.
@tprrt, could you please test this change series?

[jenswikl]

It would be nice with an alternative to interrupt_alloc_add_handler() which also takes irq_type and irq_prio into account. I wouldn't mind an interrupt_alloc_add_handler2() to take care of that, or interrupt_alloc_add_conf_handler() if you don't mind even longer function names.

[jenswikl]

For the commit "core: notif: upgrade to new interrupt framework"
s/has old/as the old/
with that fixed:
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

For the commits:

• "drivers: atmel_wdt: upgrade to new interrupt framework"
• "drivers: crypto: caam: upgrade to new interrupt framework"
• "plat-vexpress: upgrade to new interrupt framework"
• "plat-synquacer: upgrade to new interrupt framework"
• "plat-stm32mp1: upgrade to new interrupt framework"
• "core: interrupt: remove old API functions"

Please apply:
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

[etienne-lms]

Comment for adding an interrupt_alloc_add_conf_handler() helper function addressed.

[etienne-lms]

Fixed commit message (#6113 (comment)) and applied review tags to relevant patches (8/11).

[tprrt]

Hello,

Fixup commits squashed and commits reordered. @tprrt, could you please test this change series?

I just tested there is a boot issue on sama5d27, with 3f99a0fa3a1d:

E/TC:0 0 assertion '!res' failed at core/drivers/clk/clk_dt.c:161 <parse_assigned_clock>
E/TC:0 0 Panic at core/kernel/assert.c:28 <_assert_break>
E/TC:0 0 TEE load address @ 0x20000000
E/TC:0 0 Call stack:
E/TC:0 0  0x20005b89 print_kernel_stack at core/arch/arm/kernel/unwind_arm32.c:109
E/TC:0 0  0x2001bfc5 __do_panic at core/kernel/panic.c:31
E/TC:0 0  0x20017ef9 _assert_break at core/kernel/assert.c:28
E/TC:0 0  0x200595f7 _assert_trap at lib/libutils/isoc/include/assert.h:20
E/TC:0 0  0x20012cc9 parse_assigned_clock at core/drivers/clk/clk_dt.c:161 (discriminator 1)
E/TC:0 0  0x20012e61 clk_probe_assigned at core/drivers/clk/clk_dt.c:198
E/TC:0 0  0x20012e33 clk_probe_assigned at core/drivers/clk/clk_dt.c:191
E/TC:0 0  0x20012e33 clk_probe_assigned at core/drivers/clk/clk_dt.c:191
E/TC:0 0  0x20012eef clk_dt_probe at core/drivers/clk/clk_dt.c:212
E/TC:0 0  0x2001dd19 call_initcalls at core/kernel/initcall.c:41
E/TC:0 0  0x20054233 init_tee_runtime at core/arch/arm/kernel/boot.c:1379
E/TC:0 0  0x200058d5 boot_init_primary_late at core/arch/arm/kernel/boot.c:1485
E/TC:0 0  0x200001ec clear_bss at core/arch/arm/kernel/entry_a32.S:565

[etienne-lms]

Thansk @tprrt for the tests. Is the boot issue you found related to this P-R? I find it strange that clock drivers probinding fail due to changes in the interrupt framework. Can you check that sama5d27 boots well on current OP-TEE OS master branch tip?

[tprrt]

Hello

Thansk @tprrt for the tests. Is the boot issue you found related to this P-R? I find it strange that clock drivers probinding fail due to changes in the interrupt framework. Can you check that sama5d27 boots well on current OP-TEE OS master branch tip?

I agree, the issue isn’t related to this pull-request, I reproduced with the master branch (31a550c).

Kind regards,
Thomas

[etienne-lms]

@tprrt, can you bisect into master branch commits to locate (and possibly fix) the faulty commit?

[tprrt]

Hello,

@tprrt, can you bisect into master branch commits to locate (and possibly fix) the faulty commit?

Yes, I will.

[tprrt]

Hello,

Sorry, I have a very busy schedule, I will do my best to apply required changes next week.

Kind regards,
Thomas

[github-actions]

This pull request has been marked as a stale pull request because it has been open (more than) 30 days with no activity. Remove the stale label or add a comment, otherwise this pull request will automatically be closed in 5 days. Note, that you can always re-open a closed issue at any time.

[github-actions]

This pull request has been marked as a stale pull request because it has been open (more than) 30 days with no activity. Remove the stale label or add a comment, otherwise this pull request will automatically be closed in 5 days. Note, that you can always re-open a closed issue at any time.