projects / linux-2.6-block.git / blame
| Commit | Line | Data |
|---|---|---|
| d6a3b247 MCC |
1 | ================================================================= |
| 2 | CPU Scheduler implementation hints for architecture specific code | |
| 3 | ================================================================= | |
| 64c7c8f8 NP |
4 | |
| 5 | Nick Piggin, 2005 | |
| 6 | ||
| 7 | Context switch | |
| 8 | ============== | |
| 9 | 1. Runqueue locking | |
| 10 | By default, the switch_to arch function is called with the runqueue | |
| 11 | locked. This is usually not a problem unless switch_to may need to | |
| 12 | take the runqueue lock. This is usually due to a wake up operation in | |
| 95663285 | 13 | the context switch. See arch/ia64/include/asm/switch_to.h for an example. |
| 64c7c8f8 NP |
14 | |
| 15 | To request the scheduler call switch_to with the runqueue unlocked, | |
| 16 | you must `#define __ARCH_WANT_UNLOCKED_CTXSW` in a header file | |
| 17 | (typically the one where switch_to is defined). | |
| 18 | ||
| 19 | Unlocked context switches introduce only a very minor performance | |
| 20 | penalty to the core scheduler implementation in the CONFIG_SMP case. | |
| 21 | ||
| 64c7c8f8 NP |
22 | CPU idle |
| 23 | ======== | |
| 24 | Your cpu_idle routines need to obey the following rules: | |
| 25 | ||
| 26 | 1. Preempt should now disabled over idle routines. Should only | |
| 27 | be enabled to call schedule() then disabled again. | |
| 28 | ||
| 29 | 2. need_resched/TIF_NEED_RESCHED is only ever set, and will never | |
| 30 | be cleared until the running task has called schedule(). Idle | |
| 31 | threads need only ever query need_resched, and may never set or | |
| 32 | clear it. | |
| 33 | ||
| 34 | 3. When cpu_idle finds (need_resched() == 'true'), it should call | |
| 35 | schedule(). It should not call schedule() otherwise. | |
| 36 | ||
| 37 | 4. The only time interrupts need to be disabled when checking | |
| 38 | need_resched is if we are about to sleep the processor until | |
| 39 | the next interrupt (this doesn't provide any protection of | |
| d6a3b247 MCC |
40 | need_resched, it prevents losing an interrupt): |
| 41 | ||
| 42 | 4a. Common problem with this type of sleep appears to be:: | |
| 64c7c8f8 | 43 | |
| 64c7c8f8 NP |
44 | local_irq_disable(); |
| 45 | if (!need_resched()) { | |
| 46 | local_irq_enable(); | |
| 47 | *** resched interrupt arrives here *** | |
| 48 | __asm__("sleep until next interrupt"); | |
| 49 | } | |
| 50 | ||
| 51 | 5. TIF_POLLING_NRFLAG can be set by idle routines that do not | |
| 52 | need an interrupt to wake them up when need_resched goes high. | |
| 53 | In other words, they must be periodically polling need_resched, | |
| 54 | although it may be reasonable to do some background work or enter | |
| 55 | a low CPU priority. | |
| 56 | ||
| d6a3b247 MCC |
57 | - 5a. If TIF_POLLING_NRFLAG is set, and we do decide to enter |
| 58 | an interrupt sleep, it needs to be cleared then a memory | |
| 59 | barrier issued (followed by a test of need_resched with | |
| 60 | interrupts disabled, as explained in 3). | |
| 64c7c8f8 | 61 | |
| 25eb650a | 62 | arch/x86/kernel/process.c has examples of both polling and |
| 64c7c8f8 NP |
63 | sleeping idle functions. |
| 64 | ||
| 65 | ||
| 66 | Possible arch/ problems | |
| 67 | ======================= | |
| 68 | ||
| 69 | Possible arch problems I found (and either tried to fix or didn't): | |
| 70 | ||
| 64c7c8f8 NP |
71 | ia64 - is safe_halt call racy vs interrupts? (does it sleep?) (See #4a) |
| 72 | ||
| 73 | sh64 - Is sleeping racy vs interrupts? (See #4a) | |
| 74 | ||
| 75 | sparc - IRQs on at this point(?), change local_irq_save to _disable. | |
| 76 | - TODO: needs secondary CPUs to disable preempt (See #1) |