Commit graph

475 commits

Author SHA1 Message Date
Nick Piggin
7835e98b2e [PATCH] remove set_page_count() outside mm/
set_page_count usage outside mm/ is limited to setting the refcount to 1.
Remove set_page_count from outside mm/, and replace those users with
init_page_count() and set_page_refcounted().

This allows more debug checking, and tighter control on how code is allowed
to play around with page->_count.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 07:54:02 -08:00
David S. Miller
dcc1e8dd88 [SPARC64]: Add a secondary TSB for hugepage mappings.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-22 01:15:14 -08:00
David S. Miller
14778d9072 [SPARC]: Respect vm_page_prot in io_remap_page_range().
Make sure the callers do a pgprot_noncached() on
vma->vm_page_prot.

Pointed out by Hugh Dickens.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-22 01:15:13 -08:00
David S. Miller
ac0eb3eb7e [SPARC64]: Update defconfig.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:23:43 -08:00
David S. Miller
f6b83f070e [SPARC64]: Fix 2 bugs in huge page support.
1) huge_pte_offset() did not check the page table hierarchy
   elements as being empty correctly, resulting in an OOPS

2) Need platform specific hugetlb_get_unmapped_area() to handle
   the top-down vs. bottom-up address space allocation strategies.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:17:17 -08:00
Andrew Morton
467418f350 [SPARC64]: CONFIG_BLK_DEV_RAM fix
init/do_mounts_rd.c depends upon CONFIG_BLK_DEV_RAM, not CONFIG_BLK_DEV_INITRD.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:41 -08:00
David S. Miller
bb8646d834 [SPARC64]: Optimized TSB table initialization.
We only need to write an invalid tag every 16 bytes,
so taking advantage of this can save many instructions
compared to the simple memset() call we make now.

A prefetching implementation is implemented for sun4u
and a block-init store version if implemented for Niagara.

The next trick is to be able to perform an init and
a copy_tsb() in parallel when growing a TSB table.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:41 -08:00
David S. Miller
88d7079458 [SPARC64]: Allow CONFIG_MEMORY_HOTPLUG to build.
online_page() is straightforward, and then add a dummy
remove_memory() that returns -EINVAL just like i386.

There is no point in implementing remove_memory() since
__remove_pages() has no implementation either.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:40 -08:00
David S. Miller
9b4006dcf6 [SPARC64]: Use SLAB caches for TSB tables.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:39 -08:00
David S. Miller
b52439c22c [SPARC64]: Don't kill the page allocator when growing a TSB.
Try only lightly on > 1 order allocations.

If a grow fails, we are under memory pressure, so do not try
to grow the TSB for this address space any more.

If a > 0 order TSB allocation fails on a new fork, retry using
a 0 order allocation.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:38 -08:00
David S. Miller
05f9ca8359 [SPARC64]: Randomize mm->mmap_base when PF_RANDOMIZE is set.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:37 -08:00
David S. Miller
d61e16df94 [SPARC64]: Increase top of 32-bit process stack.
Put it one page below the top of the 32-bit address space.
This gives us ~16MB more address space to work with.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:36 -08:00
David S. Miller
a91690ddd0 [SPARC64]: Top-down address space allocation for 32-bit tasks.
Currently allocations are very constrained for 32-bit processes.
It grows down-up from 0x70000000 to 0xf0000000 which gives about
2GB of stack + dynamic mmap() space.

So support the top-down method, and we need to override the
generic helper function in order to deal with D-cache coloring.

With these changes I was able to squeeze out a mmap() just over
3.6GB in size in a 32-bit process.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:35 -08:00
David S. Miller
7a1ac52641 [SPARC64]: Fix and re-enable dynamic TSB sizing.
This is good for up to %50 performance improvement of some test cases.
The problem has been the race conditions, and hopefully I've plugged
them all up here.

1) There was a serious race in switch_mm() wrt. lazy TLB
   switching to and from kernel threads.

   We could erroneously skip a tsb_context_switch() and thus
   use a stale TSB across a TSB grow event.

   There is a big comment now in that function describing
   exactly how it can happen.

2) All code paths that do something with the TSB need to be
   guarded with the mm->context.lock spinlock.  This makes
   page table flushing paths properly synchronize with both
   TSB growing and TLB context changes.

3) TSB growing events are moved to the end of successful fault
   processing.  Previously it was in update_mmu_cache() but
   that is deadlock prone.  At the end of do_sparc64_fault()
   we hold no spinlocks that could deadlock the TSB grow
   sequence.  We also have dropped the address space semaphore.

While we're here, add prefetching to the copy_tsb() routine
and put it in assembler into the tsb.S file.  This piece of
code is quite time critical.

There are some small negative side effects to this code which
can be improved upon.  In particular we grab the mm->context.lock
even for the tsb insert done by update_mmu_cache() now and that's
a bit excessive.  We can get rid of that locking, and the same
lock taking in flush_tsb_user(), by disabling PSTATE_IE around
the whole operation including the capturing of the tsb pointer
and tsb_nentries value.  That would work because anyone growing
the TSB won't free up the old TSB until all cpus respond to the
TSB change cross call.

I'm not quite so confident in that optimization to put it in
right now, but eventually we might be able to and the description
is here for reference.

This code seems very solid now.  It passes several parallel GCC
bootstrap builds, and our favorite "nut cruncher" stress test which is
a full "make -j8192" build of a "make allmodconfig" kernel.  That puts
about 256 processes on each cpu's run queue, makes lots of process cpu
migrations occur, causes lots of page table and TLB flushing activity,
incurs many context version number changes, and it swaps the machine
real far out to disk even though there is 16GB of ram on this test
system. :-)

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:16:33 -08:00
David S. Miller
0c51ed93ca [SPARC64]: First cut at VIS simulator for Niagara.
Niagara does not implement some of the VIS instructions in
hardware, so we have to emulate them.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:26 -08:00
David S. Miller
90a6646bf6 [SPARC64]: Fix system type in /proc/cpuinfo and remove bogus OBP check.
Report 'sun4v' when appropriate in /proc/cpuinfo

Remove all the verifications of the OBP version string.  Just
make sure it's there, and report it raw in the bootup logs and
via /proc/cpuinfo.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:25 -08:00
David S. Miller
8935dced54 [SPARC64]: Add SMT scheduling support for Niagara.
The mapping is a simple "(cpuid >> 2) == core" for now.
Later we'll add more sophisticated code that will walk
the sun4v machine description and figure this out from
there.

We should also add core mappings for jaguar and panther
processors.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:24 -08:00
David S. Miller
17b0e199a1 [SPARC64]: Fix 32-bit truncation which broke sparsemem.
The page->flags manipulations done by the D-cache dirty
state tracking was broken because the constants were not
marked with "UL" to make them 64-bit, which means we were
clobbering the upper 32-bits of page->flags all the time.

This doesn't jive well with sparsemem which stores the
section and indexing information in the top 32-bits of
page->flags.

This is yet another sparc64 bug which has been with us
forever.

While we're here, tidy up some things in bootmem_init()
and paginig_init():

1) Pass min_low_pfn to init_bootmem_node(), it's identical
   to (phys_base >> PAGE_SHIFT) but we should use consistent
   with the variable names we print in CONFIG_BOOTMEM_DEBUG

2) max_mapnr, although no longer used, was being set
   inaccurately, we shouldn't subtract pfn_base any more.

3) All the games with phys_base in the zones_*[] arrays
   we pass to free_area_init_node() are no longer necessary.

Thanks to Josh Grebe and Fabbione for the bug reports
and testing.  Fix also verified locally on an SB2500
which had a memory layout that triggered the same problem.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:23 -08:00
David S. Miller
d1112018b4 [SPARC64]: Move over to sparsemem.
This has been pending for a long time, and the fact
that we waste a ton of ram on some configurations
kind of pushed things over the edge.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:22 -08:00
David S. Miller
ee29074d3b [SPARC64]: Fix new context version SMP handling.
Don't piggy back the SMP receive signal code to do the
context version change handling.

Instead allocate another fixed PIL number for this
asynchronous cross-call.  We can't use smp_call_function()
because this thing is invoked with interrupts disabled
and a few spinlocks held.

Also, fix smp_call_function_mask() to count "cpus" correctly.
There is no guarentee that the local cpu is in the mask
yet that is exactly what this code was assuming.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:21 -08:00
David S. Miller
a77754b4d0 [SPARC64]: Bulletproof MMU context locking.
1) Always spin_lock_init() in init_context().  The caller essentially
   clears it out, or copies the mm info from the parent.  In both
   cases we need to explicitly initialize the spinlock.

2) Always do explicit IRQ disabling while taking mm->context.lock
   and ctx_alloc_lock.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:20 -08:00
Eric Sesterhenn
9132983ae1 [SPARC64]: kzalloc() conversion
this patch converts arch/sparc64 to kzalloc usage.
Crosscompile tested with allyesconfig.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:19 -08:00
David S. Miller
f7c00338cf [SPARC64]: Fix loop termination in mark_kpte_bitmap()
If we were aligned, but didn't have at least 256MB left
to process, we would loop forever.

Thanks to fabbione for the report and testing the fix.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:19 -08:00
David S. Miller
74ae998772 [SPARC64]: Simplify TSB insert checks.
Don't try to avoid putting non-base page sized entries
into the user TSB.  It actually costs us more to check
this than it helps.

Eventually we'll have a multiple TSB scheme for user
processes.  Once a process starts using larger pages,
we'll allocate and use such a TSB.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:18 -08:00
David S. Miller
3cab0c3e86 [SPARC64]: More SUN4V cpu mondo bug fixing.
This cpu mondo sending interface isn't all that easy to
use correctly...

We were clearing out the wrong bits from the "mask" after getting
something other than EOK from the hypervisor.

It turns out the hypervisor can just be resent the same cpu_list[]
array, with the 0xffff "done" entries still in there, and it will do
the right thing.

So don't update or try to rebuild the cpu_list[] array to condense it.

This requires the "forward_progress" check to be done slightly
differently, but this new scheme is less bug prone than what we were
doing before.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:17 -08:00
David S. Miller
bcc28ee0bf [SPARC64]: Fix sun4v mna winfixup handling.
We were clobbering a base register before we were done
using it.  Fix a comment typo while we're here.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:16 -08:00
David S. Miller
c4f8ef77f9 [SPARC64]: Fix mini RTC driver reading.
Need to subtract 1900 from year and 1 from month before
giving it back to userspace.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:15 -08:00
David S. Miller
8bcd174116 [SPARC64]: Do not allow mapping pages within 4GB of 64-bit VA hole.
The UltraSPARC T1 manual recommends this because the chip
could instruction prefetch into the VA hole, and this would
also make decoding  certain kinds of memory access traps
more difficult (because the chip sign extends certain pieces
of trap state).

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:14 -08:00
David S. Miller
45f791eb0f [SPARC64]: Fix _PAGE_EXEC handling.
First of all, use the known _PAGE_EXEC_{4U,4V} value instead
of loading _PAGE_EXEC from memory.  We either know which one
to use by context, or we can code patch the test.

Next, we need to check executability of a PTE in the generic
TSB miss handler.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:13 -08:00
David S. Miller
92daa77e9a [SPARC64]: Fix typo in SUN4V D-TLB miss handler.
Should put FAULT_CODE_DTLB into %g3 not FAULT_CODE_ITLB.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:12 -08:00
David S. Miller
8ba706a95b [SPARC64]: Add mini-RTC driver for Starfire and SUN4V.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:10 -08:00
David S. Miller
b830ab665a [SPARC64]: Fix bugs in SUN4V cpu mondo dispatch.
There were several bugs in the SUN4V cpu mondo dispatch code.

In fact, if we ever got a EWOULDBLOCK or other error from
the hypervisor call, we'd potentially send a cpu mondo multiple
times to the same cpu and even worse we could loop until the
timeout resending the same mondo over and over to such cpus.

So let's bulletproof this thing as follows:

1) Implement cpu_mondo_send() and cpu_state() hypervisor calls
   in arch/sparc64/kernel/entry.S, add prototypes to asm/hypervisor.h

2) Don't build and update the cpulist using inline functions, this
   was causing the cpu mask to not get updated in the caller.

3) Disable interrupts during the entire mondo send, otherwise our
   cpu list and/or mondo block could get overwritten if we take
   an interrupt and do a cpu mondo send on the current cpu.

4) Check for all possible error return types from the cpu_mondo_send()
   hypervisor call.  In particular:

   HV_EOK) Our work is done, all cpus have received the mondo.
   HV_CPUERROR) One or more of the cpus in the cpu list we passed
                to the hypervisor are in error state.  Use cpu_state()
                calls over the entries in the cpu list to see which
		ones.  Record them in "error_mask" and report this
		after we are done sending the mondo to cpus which are
		not in error state.
   HV_EWOULDBLOCK) We need to keep trying.

   Any other error we consider fatal, we report the event and exit
   immediately.

5) We only timeout if forward progress is not made.  Forward progress
   is defined as having at least one cpu get the mondo successfully
   in a given cpu_mondo_send() call.  Otherwise we bump a counter
   and delay a little.  If the counter hits a limit, we signal an
   error and report the event.

Also, smp_call_function_mask() error handling reports the number
of cpus incorrectly.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:09 -08:00
David S. Miller
aac0aadf09 [SPARC64]: Fix bugs in SMP TLB context version expiration handling.
1) We must flush the TLB, duh.

2) Even if the sw context was seen to be valid, the local cpu's
   hw context can be out of date, so reload it unconditionally.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:08 -08:00
David S. Miller
6889331a12 [SPARC64]: Fix indexing into kpte_linear_bitmap.
Need to shift back up by 3 bits to get 8-byte entry
index.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:07 -08:00
David S. Miller
7a591cfe4e [SPARC64]: Avoid dcache-dirty page state management on sun4v.
It is totally wasted work, since we have no D-cache aliasing
issues on sun4v.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:06 -08:00
David S. Miller
2a3a5f5ddb [SPARC64]: Bulletproof hypervisor TLB flushing.
Check TLB flush hypervisor calls for errors and report them.

Pass HV_MMU_ALL always for now, we can add back the optimization
to avoid the I-TLB flush later.

Always explicitly page align the virtual address arguments.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:05 -08:00
David S. Miller
6cc80cfab8 [SPARC64]: Report mondo error correctly in hypervisor_xcall_deliver().
It's in "arg0" not "func".

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:04 -08:00
David S. Miller
3634476239 [SPARC64]: Niagara optimized XOR functions for RAID.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:03 -08:00
Andrew Morton
c4e9249b19 [SPARC64]: Fix binfmt_aout32.c build.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:02 -08:00
David S. Miller
77b838fa1e [SPARC64]: destroy_context() needs to disable interrupts.
get_new_mmu_context() can be invoked from interrupt context
now for the new SMP version wrap handling.

So disable interrupt while taking ctx_alloc_lock in destroy_context()
so we don't deadlock.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:01 -08:00
David S. Miller
a0663a79ad [SPARC64]: Fix TLB context allocation with SMT style shared TLBs.
The context allocation scheme we use depends upon there being a 1<-->1
mapping from cpu to physical TLB for correctness.  Chips like Niagara
break this assumption.

So what we do is notify all cpus with a cross call when the context
version number changes, and if necessary this makes them allocate
a valid context for the address space they are running at the time.

Stress tested with make -j1024, make -j2048, and make -j4096 kernel
builds on a 32-strand, 8 core, T2000 with 16GB of ram.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:14:00 -08:00
David S. Miller
074d82cf68 [SPARC64]: Put syscall tables after trap table.
Otherwise with too much stuff enabled in the kernel config
we can end up with an unaligned trap table.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:59 -08:00
David S. Miller
b2bef4424c [SPARC64]: Export _PAGE_E and _PAGE_CACHE to modules.
SBUS flash driver needs it.

Noticed by Fabbione.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:58 -08:00
David S. Miller
fc50492867 [SPARC64]: Drop %gl to 0 before re-enabling PSTATE_IE in rtrap
If we take a window fault, on SUN4V set %gl to zero before we
turn PSTATE_IE back on in %pstate.  Otherwise if we take an
interrupt we'll end up with corrupt register state.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:57 -08:00
David S. Miller
d7744a0950 [SPARC64]: Create a seperate kernel TSB for 4MB/256MB mappings.
It can map all of the linear kernel mappings with zero TSB hash
conflicts for systems with 16GB or less ram.  In such cases, on
SUN4V, once we load up this TSB the first time with all the
mappings, we never take a linear kernel mapping TLB miss ever
again, the hypervisor handles them all.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:56 -08:00
David S. Miller
9cc3a1ac9a [SPARC64]: Make use of Niagara 256MB PTEs for kernel mappings.
We use a bitmap, one bit for every 256MB of memory.  If the
bit is set we can use a 256MB PTE for linear mappings, else
we have to use a 4MB PTE.

SUN4V support is there, and we can very easily add support
for Panther cpu 256MB PTEs in the future.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:55 -08:00
David S. Miller
30c91d576e [SPARC64]: Use sun4v_cpu_idle() in cpu_idle() on SUN4V.
We have to turn off the "polling nrflag" bit when we sleep
the cpu like this, so that we'll get a cross-cpu interrupt
to wake the processor up from the yield.

We also have to disable PSTATE_IE in %pstate around the yield
call and recheck need_resched() in order to avoid any races.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:54 -08:00
David S. Miller
689126a48a [SPARC64] math-emu: Delete debugging printk left by previous commit.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:53 -08:00
David S. Miller
6f5374c91f [SPARC64]: Add sun4v_cpu_yield().
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:52 -08:00
David S. Miller
1bd0cd74d1 [SPARC64]: Kill cpudata->idle_volume.
Set, but never used.

We used to use this for dynamic IRQ retargetting, but that
code died a long time ago.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:51 -08:00
David S. Miller
8ca2557c48 [SPARC64]: Niagara optimized memset/bzero/clear_user.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:50 -08:00
David S. Miller
d371c0c174 [SPARC64]: Pass multiple CPUs at once to hypervisor cross-call API.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:49 -08:00
David S. Miller
c79f76777d [SPARC64]: Args to SUNW,set-trap-table are 64-bit.
They were getting truncated to 32-bit and this is very bad
when your MMU fault status area is in physical memory above
4GB on SUN4V.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:48 -08:00
David S. Miller
4e74ae800b [SPARC64]: Handle unimplemented FPU square-root on Niagara.
The math-emu code only expects unfinished fpop traps when
emulating FPU sqrt instructions on pre-Niagara chips.
On Niagara we can get unimplemented fpop, so handle that.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:48 -08:00
David S. Miller
55555633bd [SPARC64]: Typo in sun4v_data_access_exception log message.
Should be "Dax" not "Iax".

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:46 -08:00
David S. Miller
d82965c167 [SPARC64]: Handle zero-length map requests in pci_sun4v.c
By simply changing the do-while loop into a plain
while loop.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:45 -08:00
David S. Miller
abf3b7bd89 [SPARC64]: Kill stray PGLIST_NENTS check in pci_sun4v.c
I forgot to remove the one in pci_4v_map_sg() during the
iommu batching commit.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:44 -08:00
David S. Miller
39334a4b2c [SPARC64]: Fix typo in dump_tl1_traplog()
Actually make use of the 'limit' we compute.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:43 -08:00
David S. Miller
37133c006c [SPARC64]: Disable smp_report_regs() for now.
It's extremely noisy and causes much grief on slow
consoles with large numbers of cpus.

We'll have to provide this some saner way in order
to re-enable this.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:42 -08:00
David S. Miller
6a32fd4d0d [SPARC64]: Remove PGLIST_NENTS PCI IOMMU mapping limitation on SUN4V.
Use a batching queue system for IOMMU mapping setup,
with a page sized batch.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:41 -08:00
David S. Miller
04d74758eb [SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors.
We're about to seriously die in these cases so it is important
that the messages make it to the console.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:40 -08:00
David S. Miller
24c523ecc6 [SPARC64]: Fix unaligned access winfxup handling on SUN4V.
Another case where we have to force ourselves into global register
level one.  Also make sure the arguments passed to sun4v_do_mna() are
correct.

This area actually needs some more work, for example spill fixup is
not necessarily going to do the right thing for this case.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:39 -08:00
David S. Miller
6cc200db95 [SPARC64]: Set %gl to 1 in kvmap_itlb_longpath on SUN4V.
Just like kvmap_dtlb_longpath we have to force the
global register level to one in order to mimick the
PSTATE_MG --> PSTATE_AG trasition done on SUN4U.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:39 -08:00
David S. Miller
0f15952ac8 [SPARC64]: Export a PAGE_SHARED symbol.
For drivers/media/*, noticed by Fabbione.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:36 -08:00
David S. Miller
8b23427441 [SPARC64]: More TLB/TSB handling fixes.
The SUN4V convention with non-shared TSBs is that the context
bit of the TAG is clear.  So we have to choose an "invalid"
bit and initialize new TSBs appropriately.  Otherwise a zero
TAG looks "valid".

Make sure, for the window fixup cases, that we use the right
global registers and that we don't potentially trample on
the live global registers in etrap/rtrap handling (%g2 and
%g6) and that we put the missing virtual address properly
in %g5.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:34 -08:00
David S. Miller
7adb37fe80 [SPARC64]: Don't do anything in flush_ptrace_access() on SUN4V.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:33 -08:00
David S. Miller
6c8927c963 [SPARC64]: Fix some SUN4V TLB handling bugs.
1) Add error return checking for TLB load hypervisor
   calls.

2) Don't fallthru to dtlb tsb miss handler from itlb tsb
   miss handler, oops.

3) On window fixups, propagate fault information to fixup
   handler correctly.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:32 -08:00
David S. Miller
12e126ad22 [SPARC64]: Check for errors in hypervisor_tlb_lock().
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:31 -08:00
David S. Miller
52845cdb3b [SPARC64]: Init boot cpu's trap_block[] before paging_init()
It must be ready when we take over the trap table.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:30 -08:00
David S. Miller
3763be32d5 [SPARC64]: Define ARCH_HAS_READ_CURRENT_TIMER.
This gives more consistent bogomips and delay() semantics,
especially on sun4v.  It gives weird looking values though...

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:29 -08:00
David S. Miller
3f19a84e39 [SPARC64]: Set associativity of kernel TSB descriptor correctly.
It should be 1, not 0.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:29 -08:00
David S. Miller
c857e3fdbc [SPARC64]: __bzero_noasi --> __clear_user
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:28 -08:00
David S. Miller
46f8604714 [SPARC64]: Put SUN4V ITSB miss into correct trap table entry.
It's 0x9 not 0xb.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:27 -08:00
David S. Miller
3b3ab2eb9c [SPARC64]: Use phys tsb address in tsb_insert() in SUN4V.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:25 -08:00
David S. Miller
ebd8c56c5a [SPARC64]: Fix uniprocessor IRQ targetting on SUN4V.
We need to use the real hardware processor ID when
targetting interrupts, not the "define to 0" thing
the uniprocessor build gives us.

Also, fill in the Node-ID and Agent-ID fields properly
on sun4u/Safari.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:24 -08:00
David S. Miller
101d5c18a9 [SPARC64]: Fix PCI IRQ probing regression.
If the top-level cnode had multi entries in it's "reg"
property, we'd fail.  The buffer wasn't large enough in
such cases.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:23 -08:00
David S. Miller
72aff53f1f [SPARC64]: Get SUN4V SMP working.
The sibling cpu bringup is extremely fragile.  We can only
perform the most basic calls until we take over the trap
table from the firmware/hypervisor on the new cpu.

This means no accesses to %g4, %g5, %g6 since those can't be
TLB translated without our trap handlers.

In order to achieve this:

1) Change sun4v_init_mondo_queues() so that it can operate in
   several modes.

   It can allocate the queues, or install them in the current
   processor, or both.

   The boot cpu does both in it's call early on.

   Later, the boot cpu allocates the sibling cpu queue, starts
   the sibling cpu, then the sibling cpu loads them in.

2) init_cur_cpu_trap() is changed to take the current_thread_info()
   as an argument instead of reading %g6 directly on the current
   cpu.

3) Create a trampoline stack for the sibling cpus.  We do our basic
   kernel calls using this stack, which is locked into the kernel
   image, then go to our proper thread stack after taking over the
   trap table.

4) While we are in this delicate startup state, we put 0xdeadbeef
   into %g4/%g5/%g6 in order to catch accidental accesses.

5) On the final prom_set_trap_table*() call, we put &init_thread_union
   into %g6.  This is a hack to make prom_world(0) work.  All that
   wants to do is restore the %asi register using
   get_thread_current_ds().

Longer term we should just do the OBP calls to set the trap table by
hand just like we do for everything else.  This would avoid that silly
prom_world(0) issue, then we can remove the init_thread_union hack.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:22 -08:00
David S. Miller
19a0d585e8 [SPARC64]: Disable smp_report_regs() for now.
For 32 cpus and a slow console, it just wedges the
machine especially with DETECT_SOFTLOCKUP enabled.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:21 -08:00
David S. Miller
6154f94f0e [SPARC64]: Rewrite pci_intmap_match().
The whole algorithm was wrong.  What we need to do is:

1) Walk each PCI bus above this device on the path to the
   PCI controller nexus, and for each:
      a) If interrupt-map exists, apply it, record IRQ controller node
      b) Else, swivel interrupt number using PCI_SLOT(), use PCI bus
	 parent OBP node as controller node
      c) Walk up to "controller node" until we hit the first PCI bus
	 in this domain, or "controller node" is the PCI controller
	 OBP node
2) If we walked to PCI controller OBP node, we're done.
3) Else, apply PCI controller interrupt-map to interrupt.

There is some stuff that needs to be checked out for ebus and
isa, but the PCI part is good to go.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:20 -08:00
David S. Miller
14f6689cbb [SPARC64]: Don't set interrupt state to IDLE in enable_irq().
We'll lose events that way.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:20 -08:00
David S. Miller
af02bec662 [SPARC64]: Fix return from trap on SUN4V.
We need to set the global register set _AND_ disable
PSTATE_IE in %pstate.  The original patch sequence was
leaving PSTATE_IE enabled when returning to kernel mode,
oops.

This fixes the random register corruption being seen
on SUN4V.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:19 -08:00
David S. Miller
22780e23c6 [SPARC64]: Set dummy bucket->{imap,iclr} unique on SUN4V.
So that free_irq() disable's the IRQ correctly.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:17 -08:00
David S. Miller
94f8762db9 [SPARC64]: Add sun4v_cpu_qconf() hypervisor call.
Call it from register_one_mondo().

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:16 -08:00
David S. Miller
8e42550c68 [SPARC64]: do_fptrap needs to load the thread reg into %g6.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:14 -08:00
David S. Miller
9b6b46470c [SPARC64]: Fix bogus call to sun4v_mna in winfixup code.
The C function is named sun4v_do_mna not sun4v_mna.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:13 -08:00
David S. Miller
3d6395cb77 [SPARC64]: Fix tl1 trap state capture/dump on SUN4V.
No trap levels above 2 in privileged mode on SUN4V.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:12 -08:00
David S. Miller
e7a0453ef8 [SPARC64] PCI: Size TSB correctly on SUN4V.
Forgot to multiply by 8 * 1024, oops.  Correct the size constant when
the virtual-dma arena is 2GB in size, it should bet 256 not 128.

Finally, log some info about the TSB at probe time.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:10 -08:00
David S. Miller
c7f81d42d3 [SPARC64]: Don't use ASI_QUAD_LDD_PHYS on SUN4V.
Need to use ASI_QUAD_LDD_PHYS_4V instead.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:09 -08:00
David S. Miller
a7b31bac69 [SPARC64]: Do not write garbage into %pstate in tsb_context_switch().
For SUN4V, we were clobbering %o5 to do the hypervisor call.
This clobbers the saved %pstate value and we end up writing
garbage into that register as a result.  Oops.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:08 -08:00
David S. Miller
de635d833f [SPARC64]: Fix flush_tsb_user() on SUN4V.
Needs to use physical addressing just like cheetah_plus.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:07 -08:00
David S. Miller
1daef08a12 [SPARC64]: Fix comment typo in __flush_tlb_kernel_range.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:06 -08:00
David S. Miller
9d29a3fafd [SPARC64]: Decode virtual-devices interrupts correctly.
Need to translate through the interrupt-map{,-mask] properties.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:05 -08:00
David S. Miller
7890f794e0 [SPARC64]: Add prom_{start,stop}cpu_cpuid().
Use prom_startcpu_cpuid() on SUN4V instead of prom_startcpu().

We should really test for "SUNW,start-cpu-by-cpuid" presence
and use it if present even on SUN4U.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:04 -08:00
David S. Miller
63c2a0e598 [SPARC64]: Fix pci_intmap_match().
When crawling up the PCI bus chain, stop at the first node
that has an interrupt-map property before we hit the root.

Also, if we use a bus interrupt-{map,mask} do not forget to
update the 'intmask' pointer as we do for the 'intmap' pointer.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:03 -08:00
David S. Miller
ab66a50e31 [SPARC64]: Two IRQ handling fixes.
On SUN4V, force IRQ state to idle in enable_irq().  However,
I'm still not sure this is %100 correct.

Call add_interrupt_randomness() on SUN4V too.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:02 -08:00
David S. Miller
f03b8a5468 [SPARC64]: Use different cache sizing defaults on SUN4V.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:01 -08:00
David S. Miller
329c68b218 [SPARC64]: Make lack of interrupt-map-* a fatal error on SUN4V.
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:13:00 -08:00
David S. Miller
abd92b2d21 [SPARC64]: Fix sun4v_intr_setenabled() return value check in enable_irq().
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:12:59 -08:00
David S. Miller
355db99860 [SPARC64]: Explicitly init *nregs to 0 in find_device_prom_node().
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:12:58 -08:00
David S. Miller
987b6de710 [SPARC64]: Restrict PCI bus scanning on SUN4V.
On the PBM's first bus number, only allow device 0, function 0, to be
poked at with PCI config space accesses.

For some reason, this single device responds to all device numbers.

Also, reduce the verbiage of the debugging log printk's for PCI cfg
space accesses in the SUN4V PCI controller driver, so that it doesn't
overwhelm the slow SUN4V hypervisor console.

Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-20 01:12:57 -08:00