kernel.git/arch/powerpc/kernel/smp.c, branch linux-2.6.25.y

POWERPC Bolt in SLB entry for kernel stack on secondary cpus

2008-06-09T18:26:58Z

This is upstream as commit 3b5750644b2ffa2a76fdfe7b4e00e4af2ecf3539. This fixes a regression reported by Kamalesh Bulabel where a POWER4 machine would crash because of an SLB miss at a point where the SLB miss exception was unrecoverable. This regression is tracked at: http://bugzilla.kernel.org/show_bug.cgi?id=10082 SLB misses at such points shouldn't happen because the kernel stack is the only memory accessed other than things in the first segment of the linear mapping (which is mapped at all times by entry 0 of the SLB). The context switch code ensures that SLB entry 2 covers the kernel stack, if it is not already covered by entry 0. None of entries 0 to 2 are ever replaced by the SLB miss handler. Where this went wrong is that the context switch code assumes it doesn't have to write to SLB entry 2 if the new kernel stack is in the same segment as the old kernel stack, since entry 2 should already be correct. However, when we start up a secondary cpu, it calls slb_initialize, which doesn't set up entry 2. This is correct for the boot cpu, where we will be using a stack in the kernel BSS at this point (i.e. init_thread_union), but not necessarily for secondary cpus, whose initial stack can be allocated anywhere. This doesn't cause any immediate problem since the SLB miss handler will just create an SLB entry somewhere else to cover the initial stack. In fact it's possible for the cpu to go quite a long time without SLB entry 2 being valid. Eventually, though, the entry created by the SLB miss handler will get overwritten by some other entry, and if the next access to the stack is at an unrecoverable point, we get the crash. This fixes the problem by making slb_initialize create a suitable entry for the kernel stack, if we are on a secondary cpu and the stack isn't covered by SLB entry 0. This requires initializing the get_paca()->kstack field earlier, so I do that in smp_create_idle where the current field is initialized. This also abstracts a bit of the computation that mk_esid_data in slb.c does so that it can be used in slb_initialize. Signed-off-by: Paul Mackerras Signed-off-by: Greg Kroah-Hartman Signed-off-by: Chris Wright

[POWERPC] Make smp_send_stop() handle panic and xmon reboot

2008-01-25T11:52:50Z

smp_send_stop() will send an IPI to all other cpus to shut them down. However, for the case of xmon-based reboots (as well as potentially some panics), the other cpus are (or might be) spinning with interrupts off, and won't take the IPI. Current code will drop us into the debugger when the IPI fails, which means we're in an infinite loop that we can't get out of without an external reset of some sort. Instead, make the smp_send_stop() IPI call path just print the warning about being unable to send IPIs, but make it return so the rest of the shutdown sequence can continue. It's not perfect, but the lesser of two evils. Also move the call_lock handling outside of smp_call_function_map so we can avoid deadlocks in smp_send_stop(). Signed-off-by: Olof Johansson Signed-off-by: Paul Mackerras

[POWERPC] Make smp_call_function_map static

2008-01-25T11:52:49Z

smp_call_function_map should be static, and for consistency prepend it with __ like other local helper functions in the same file. Signed-off-by: Olof Johansson Signed-off-by: Paul Mackerras

Convert cpu_sibling_map to be a per cpu variable

2007-10-16T16:42:50Z

Convert cpu_sibling_map from a static array sized by NR_CPUS to a per_cpu variable. This saves sizeof(cpumask_t) * NR unused cpus. Access is mostly from startup and CPU HOTPLUG functions. Signed-off-by: Mike Travis Cc: Andi Kleen Cc: Christoph Lameter Cc: "Siddha, Suresh B" Cc: "David S. Miller" Cc: Paul Mackerras Cc: Benjamin Herrenschmidt Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

[POWERPC] Implement clockevents driver for powerpc

2007-10-03T05:44:34Z

This registers a clock event structure for the decrementer and turns on CONFIG_GENERIC_CLOCKEVENTS, which means that we now don't need most of timer_interrupt(), since the work is done in generic code. For secondary CPUs, their decrementer clockevent is registered when the CPU comes up (the generic code automatically removes the clockevent when the CPU goes down). Signed-off-by: Tony Breeds Signed-off-by: Paul Mackerras

[POWERPC] Avoid pointless WARN_ON(irqs_disabled()) from panic codepath

2007-09-22T04:49:22Z

> ------------[ cut here ]------------ > Badness at arch/powerpc/kernel/smp.c:202 comes when smp_call_function_map() has been called with irqs disabled, which is illegal. However, there is a special case, the panic() codepath, when we do not want to warn about this -- warning at that time is pointless anyway, and only serves to scroll away the *real* cause of the panic and distracts from the real bug. * So let's extract the WARN_ON() from smp_call_function_map() into all its callers -- smp_call_function() and smp_call_function_single() * Also, introduce another caller of smp_call_function_map(), namely __smp_call_function() (and make smp_call_function() a wrapper over this) which does *not* warn about disabled irqs * Use this __smp_call_function() from the panic codepath's smp_send_stop() We also end having to move code of smp_send_stop() below the definition of __smp_call_function(). Signed-off-by: Satyam Sharma Signed-off-by: Paul Mackerras

[POWERPC] Fix num_cpus calculation in smp_call_function_map()

2007-08-03T09:36:00Z

In smp_call_function_map(), num_cpus is set to the number of online CPUs minus one. However, if the CPU mask does not include all CPUs (except the one we're running on), the routine will hang in the first while() loop until the 8 second timeout occurs. The num_cpus should be set to the number of CPUs specified in the mask passed into the routine, after we've made any modifications to the mask. With this change, we can also get rid of the call to cpus_empty() and avoid adding another pass through the bitmask. Signed-off-by: Kevin Corry Signed-off-by: Carl Love Signed-off-by: Paul Mackerras

[POWERPC] Allow smp_call_function_single() to current cpu

2007-07-22T11:30:58Z

This removes the requirement for callers to get_cpu() to check in simple cases. i386 and x86_64 already received a similar treatment. Signed-off-by: Avi Kivity Signed-off-by: Paul Mackerras

[POWERPC] Fix smp_call_function to be preempt-safe

2007-05-22T10:20:56Z

smp_call_function_map() was not safe against preemption to another cpu: its test for removing self from map was outside the spinlock. Rearrange it a little to fix that. smp_call_function_single() was also wrong: now get_cpu() before excluding self, as other architectures do. Signed-off-by: Hugh Dickins Signed-off-by: Paul Mackerras

[POWERPC] Add smp_call_function_map and smp_call_function_single

2007-05-07T10:31:13Z

Add a new function named smp_call_function_single(). This matches a generic prototype from include/linux/smp.h. Add a function smp_call_function_map(). This is, for the most part, a rename of smp_call_function, with some added cpumask support. smp_call_function and smp_call_function_single call into smp_call_function_map. Lightly tested on 970mp (blade), power4 and power5. Signed-off-by: Will Schmidt cc: Anton Blanchard Signed-off-by: Paul Mackerras