FreeCalypso > hg > freecalypso-sw
comparison gsm-fw/sysglue/irqfiq.S @ 143:afceeeb2cba1
Our nuc-fw is destined to become gsm-fw, so I went ahead and did the big hg mv
author | Michael Spacefalcon <msokolov@ivan.Harhan.ORG> |
---|---|
date | Tue, 12 Nov 2013 05:35:48 +0000 |
parents | nuc-fw/sysglue/irqfiq.S@2c5160a9d652 |
children |
comparison
equal
deleted
inserted
replaced
142:15d5977390c2 | 143:afceeeb2cba1 |
---|---|
1 /* | |
2 * This module contains the assembly shells for IRQ and FIQ, separated | |
3 * from the architectured vectors only by some simple unconditional | |
4 * branch instructions. | |
5 * | |
6 * Note that TI's way of handling interrupts sacrifices Nucleus' ability | |
7 * to nest interrupts and minimize the IRQ-disabled window: if my (Falcon's) | |
8 * understanding is correct, TI's code leaves all further IRQs disabled | |
9 * for the full execution duration of an IRQ handler. (IRQ handlers are | |
10 * really LISRs, but TI's GSM fw does not use Nucleus' LISR framework.) | |
11 */ | |
12 | |
13 .section iram.text,"ax",%progbits | |
14 .code 32 | |
15 | |
16 .globl _INT_IRQ | |
17 _INT_IRQ: | |
18 STMDB sp!,{r0-r4} @ used to be a1-a4 | |
19 | |
20 /* | |
21 * Thanks to TI for discovering and documenting this apparent ARM7TDMI bug: | |
22 | |
23 BUG correction 1st part ------------------- | |
24 It looks like there is an issue with ARM7 IRQ masking in the CPSR register | |
25 which leads to crashes in Nucleus+ scheduler. | |
26 Basically the code below (correct as LOCKOUT = 0xC0) is used in many places by N+ but do not | |
27 prevent from having an interrupt after the execution of the third line (I mean execution, not | |
28 fetch). | |
29 MRS a1,CPSR ; Pickup current CPSR | |
30 ORR a1,a1,#LOCKOUT ; Build interrupt lockout value | |
31 MSR CPSR,a1 ; Lockout interrupts | |
32 * IRQ INTERRUPT ! * | |
33 Next instructions... | |
34 | |
35 SW workaround: | |
36 When a task is interrupted at this point an interrupted context is stored on its task and will | |
37 be resumed later on at the next instruction but to make a long story short it leads to some | |
38 problem as the OS does not expect to be interrupted there. | |
39 Further testing tends to show that the CPSR *seems* to be loaded with the proper masking value | |
40 but that the IRQ is still triggered (has been hardwarewise requested during the instruction | |
41 exectution by the ARM7 core?) | |
42 */ | |
43 | |
44 MRS a1,spsr @ check for the IRQ bug: | |
45 TST a1,#0x80 @ if the I - flag is set, | |
46 BNE IRQBUG @ then postpone execution of this IRQ | |
47 /* Bug correction 1st part end --------------- */ | |
48 | |
49 SUB r4,lr,#4 @ Save IRQ's lr (return address) | |
50 BL TCT_Interrupt_Context_Save @ Call context save routine | |
51 | |
52 BL IQ_IRQ_isr @ Call int. service routine | |
53 | |
54 /* IRQ interrupt processing is complete. Restore context- Never | |
55 returns! */ | |
56 B TCT_Interrupt_Context_Restore | |
57 | |
58 /* BUG correction 2nd part ------------------ */ | |
59 IRQBUG: LDMFD sp!,{r0-r4} @ return from interrupt | |
60 SUBS pc,r14,#4 | |
61 /* BUG correction 2nd part end -------------- */ | |
62 | |
63 .globl _INT_FIQ | |
64 _INT_FIQ: | |
65 STMDB sp!,{r0-r4} @ used to be a1-a4 | |
66 SUB r4,lr,#4 @ Save FIQ's lr (return address) | |
67 BL TCT_Interrupt_Context_Save @ Call context save routine | |
68 | |
69 BL IQ_FIQ_isr @ Call the FIQ ISR | |
70 | |
71 /* FIQ interrupt processing is complete. Restore context- Never | |
72 returns! */ | |
73 B TCT_Interrupt_Context_Restore |