ARCv2: entry: simplify return to Delay Slot via interrupt

Commit 4255b07f ("ARCv2: STAR 9000793984: Handle return
from intr to Delay Slot") involved a complex 2 staged trampoline.

Apparently this can be greatly simplified by returning from pure
kernel mode (iso interrupt) so drop to pure kernel mdoe and execute
the normal exception return path.

Testing this was a bit of challenge as return from interrupt is rarely
executed now after commit 4de0e528 ("ARCv2: STAR 9000814690:
Really Re-enable interrupts to avoid deadlocks"). That fix is necessary
evil and pct interrupts etc do exercise intr return path.
Anyhow after a revert of above in my local test setup I was able to hit
this case and verify the patch works.
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>

arch/arc/kernel/entry-arcv2.S

@@ -79,7 +79,7 @@ ENTRY(handle_interrupt)
 	#
 	# Note this disable is only for consistent book-keeping as further interrupts
 	# will be disabled anyways even w/o this. Hardware tracks active interrupts
-	# seperately in AUX_IRQ_ACTIVE.active and will not take new interrupts
+	# seperately in AUX_IRQ_ACT.active and will not take new interrupts
 	# unless this one returns (or higher prio becomes pending in 2-prio scheme)
 
 	IRQ_DISABLE
@@ -200,17 +200,18 @@ restore_regs:
 	ld	r0, [sp, PT_status32]	; U/K mode at time of entry
 	lr	r10, [AUX_IRQ_ACT]
 
-	bmsk	r11, r10, 15	; AUX_IRQ_ACT.ACTIVE
+	bmsk	r11, r10, 15		; extract AUX_IRQ_ACT.active
 	breq	r11, 0, .Lexcept_ret	; No intr active, ret from Exception
 
 ;####### Return from Intr #######
 
+.Lisr_ret:
+
 debug_marker_l1:
 	; bbit1.nt r0, STATUS_DE_BIT, .Lintr_ret_to_delay_slot
 	btst	r0, STATUS_DE_BIT		; Z flag set if bit clear
 	bnz	.Lintr_ret_to_delay_slot	; branch if STATUS_DE_BIT set
 
-.Lisr_ret_fast_path:
 	; Handle special case #1: (Entry via Exception, Return via IRQ)
 	;
 	; Exception in U mode, preempted in kernel, Intr taken (K mode), orig
@@ -244,8 +245,8 @@ debug_marker_syscall:
 ;
 ; IRQ RTIE won't reliably restore DE bit and/or BTA, needs workaround
 ;
-; Solution is return from Intr w/o any delay slot quirks into a kernel trampoline
-; and from pure kernel mode return to delay slot which handles DS bit/BTA correctly
+; Solution is to drop out of interrupt context into pure kernel mode
+; and return from pure kernel mode which does right things for delay slot
 
 .Lintr_ret_to_delay_slot:
 debug_marker_ds:
@@ -254,48 +255,9 @@ debug_marker_ds:
 	add	r2, r2, 1
 	st	r2, [@intr_to_DE_cnt]
 
-	ld	r2, [sp, PT_ret]
-	ld	r3, [sp, PT_status32]
-
-	; STAT32 for Int return created from scratch
-	; (No delay dlot, disable Further intr in trampoline)
-
-	bic  	r0, r3, STATUS_U_MASK|STATUS_DE_MASK|STATUS_IE_MASK|STATUS_L_MASK
-	st	r0, [sp, PT_status32]
-
-	mov	r1, .Lintr_ret_to_delay_slot_2
-	st	r1, [sp, PT_ret]
-
-	; Orig exception PC/STAT32 safekept @orig_r0 and @event stack slots
-	st	r2, [sp, 0]
-	st	r3, [sp, 4]
-
-	b	.Lisr_ret_fast_path
-
-.Lintr_ret_to_delay_slot_2:
-	; Trampoline to restore orig exception PC/STAT32/BTA/AUX_USER_SP
-	sub	sp, sp, SZ_PT_REGS
-	st	r9, [sp, -4]
-
-	ld	r9, [sp, 0]
-	sr	r9, [eret]
-
-	ld	r9, [sp, 4]
-	sr	r9, [erstatus]
-
-	; restore AUX_USER_SP if returning to U mode
-	bbit0	r9, STATUS_U_BIT, 1f
-	ld	r9, [sp, PT_sp]
-	sr	r9, [AUX_USER_SP]
-
-1:
-	ld	r9, [sp, 8]
-	sr	r9, [erbta]
-
-	ld	r9, [sp, -4]
-	add	sp, sp, SZ_PT_REGS
-
-	; return from pure kernel mode to delay slot
-	rtie
+	; drop out of interrupt context (clear AUX_IRQ_ACT.active)
+	bmskn	r11, r10, 15
+	sr	r11, [AUX_IRQ_ACT]
+	b	.Lexcept_ret
 
 END(ret_from_exception)