First, in huge_pte_offset(), we were erroneously checking
pgd_present(), which is always true, rather than pud_present(),
which is the thing that tells us if there is a top-level (L0) PTE.
Fixing this means we properly look up huge page entries only when
the Present bit is actually set in the PTE.

Second, use the standard pte_alloc_map() instead of the hand-rolled
pte_alloc_hugetlb() routine that basically was written to avoid
worrying about CONFIG_HIGHPTE.  However, we no longer plan to support
HIGHPTE, so a separate routine was just unnecessary code duplication.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Also, alphabetize the existing entries for tile.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change adds support for avoiding recursive backtracer crashes;
we haven't seen this in practice other than when things are seriously
corrupt, but it may help avoid losing the root cause of a crash.

Also, don't abort kernel backtracers for invalid userspace PC's.
If we do, we lose the ability to backtrace through a userspace
call to a bad address above PAGE_OFFSET, even though that it can
be perfectly reasonable to continue the backtrace in such a case.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change enables unaligned userspace memory access via a kernel
fast path on tilegx.  The kernel tracks user PC/instruction pairs
per-thread using a direct-mapped cache in userspace.  The cache
maps those PC/instruction pairs to JIT'ed instruction sequences that
load or store using byte-wide load store intructions and then
synthesize 2-, 4- or 8-byte load or store results.  Once an
instruction has been seen to generate an unaligned access once,
subsequent hits on that instruction typically require overhead
of only around 50 cycles if cache and TLB is hot.

We support the prctl() PR_GET_UNALIGN / PR_SET_UNALIGN sys call to
enable or disable unaligned fixups on a per-process basis.

To do this we pull some of the tilepro unaligned support out of the
single_step.c file; tilepro uses instruction disassembly for both
single-step and unaligned access support.  Since tilegx actually has
hardware singlestep support, though, it's cleaner to keep the tilegx
unaligned access code in a separate file.  While we're at it,
properly rename the tilepro-specific types, etc., to have tilepro
suffixes instead of generic tile suffixes.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Pointed out by checkpatch.  A few of the DEFINE() lines were
properly written without backslash continuation; fix the rest.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

As specified, the test wasn't correct, and in any case it should
be a BUILD_BUG_ON.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change adds support for the "memmap" boot parameter similar
to what x86 provides.  The tile version supports "memmap=1G$5G",
for example, as a way to reserve a 1 GB range starting at PA 5GB.
The memory is reserved via bootmem during startup, and we create a
suitable "struct resource" marked as "Reserved" so you can see the
range reported by /proc/iomem.  Up to 64 such regions can currently
be reserved on the boot command line.

We do not support the x86 options "memmap=nn@ss" (force some memory
to be available at the given address) since it's pointless to try to
have Linux use memory the Tilera hypervisor hasn't given it.  We do
not support "memmap=nn#ss" to add an ACPI range for later processing,
since we don't support ACPI.  We do not support "memmap=exactmap"
since we don't support reading the e820 information from the BIOS
like x86 does.  I did add support for "memmap=nn" (and the synonym
"mem=nn") which cap the highest PA value at "nn"; these are both
just a synonym for the existing tile boot option "maxmem".
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change improves and cleans up the tile console.

- We enable HVC_IRQ support on tilegx, with the addition of a new
  Tilera hypervisor API for tilegx to allow a console IPI.  If IPI
  support is not available we fall back to the previous polling mode.

- We simplify the earlyprintk code to use CON_BOOT and eliminate some
  of the other supporting earlyprintk code.

- A new tile_console_write() primitive is used to send output to
  the console and is factored out of the hvc_tile driver.
  This lets us support a "sim_console" boot argument to allow using
  simulator hooks to send output to the "console" as a slightly
  faster alternative to emulating the hardware more directly.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

We can take advantage of the fact that bit 29 is hard-wired
to zero in register TRIO_TILE_PIO_REGION_SETUP_CFG_ADDR.
This is handy since at the moment we only allocate one 4GB
region for vmalloc, and with this change we can allocate
four or more TRIO MACs without using up all the vmalloc space.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

On Tilera Gx72 systems, the logic for figuring out whether
a given port is root complex is slightly different.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The code incorrectly masked with PAGE_OFFSET instead of
PAGE_SIZE-1.  This only matters when trying to do a
non page-aligned DMA; it was noticed during code inspection.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The standard kernel function dma_get_required_mask() uses the
highest DRAM address to determine if 32-bit or 64-bit DMA addressing
is needed.  This only works on architectures that have direct mapping
between the PA and the PCI address space, i.e. those that don't have
I/O TLBs or have I/O TLB but choose to use direct mapping.  Neither
of these are true for tilegx.  Whether to use 64-bit DMA should depend
on the PCI device's capability only, not on the amount of DRAM
installeds, so we now advertise a 64-bit DMA mask unconditionally.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Using the low-level hv_dev_pread() API makes assumptions about the
layout of datastructures in the Tilera hypervisor API; it's better to
use the gxio_XXX accessor and the pcie_trio_ports_property struct.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change is purely stylistic but improves the readability
of the tile PCI RC driver.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The .mem_resources[] field in the pci_controller struct
is now obsoleted by the .mem_space and .io_space fields.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The TRIO shim initialization is shared with other kernel drivers
such as the endpoint and StreamIO drivers, so reorganize the
initialization flow to ensure that the root complex driver properly
initializes TRIO state regardless of what kind of TRIO driver will
end up using the shim.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

To enable this functionality, configure CONFIG_TILE_PCI_IO.  Without
this flag, the kernel still assigns I/O address ranges to the
devices, but no TRIO resource and mapping support is provided.

We assign disjoint I/O address ranges to separate PCIe domains.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Besides using pr_info() to print the linkdown status for a plug-in
slot, add extra indication that this is expected if the slot is empty.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

To support PCIe devices with higher number of MSI-X interrupt vectors,
e.g. 16 for the LSI RAID card, enhance the Gx RC stack to provide more
MSI-X vectors by using the TRIO Scatter Queues, which provide 8 more
vectors in addition to ~10 from the Map Mem regions.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The LSI MEGARAID SAS HBA suffers from the problem where it can do
64-bit DMA to streaming buffers but not to consistent buffers.
In other words, 64-bit DMA is used for disk data transfers and 32-bit
DMA must be used for control message transfers. According to LSI,
the firmware is not fully functional yet. This change implements a
kind of hybrid dma_ops to support this.

Note that on most other platforms, the 64-bit DMA addressing space is the
same as the 32-bit DMA space and they overlap the physical memory space.
No special arrangement is needed to support this kind of mixed DMA
capability.  On TILE-Gx, the 64-bit DMA space is completely separate
from the 32-bit DMA space.  Due to the use of the IOMMU, the 64-bit DMA
space doesn't overlap the physical memory space.  On the other hand,
the 32-bit DMA space overlaps the physical memory space under 4GB.
The separate address spaces make it necessary to have separate dma_ops.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

If we are rebooting (e.g. via kexec) then the PCI RC link may
already be up.  In that case, we don't want to do the software
fixup to force the link up, since that can degrade it to Gen1.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Allow longer delays if requested, and print the info messages
as we are performing the delay, not when parsing the arguments.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Fix a bug in the tilepro PCI resource allocation code that could make
the bootmem allocator unhappy if 4GB is installed on mshim 0.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

- remove unneeded <linux/bootmem.h> include in pci.c
- eliminate unused pci_controller.first_busno field
- prefer msleep to mdelay
- remove stale comment about pci_scan_bus_parented()
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Using strlen as a model, add length checking to create strnlen.
Signed-off-by: Ken Steele <ken@tilera.com>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This change cleans up the string code in a number of ways:

- For memcpy(), fix bug in prefetch and increase distance to 3 lines;
  optimize for unaligned data; do all loads before wh64 to make memcpy
  safe for forward-overlapping calls; etc.  Performance is improved.

- Use new copy_byte() function on tilegx to spread a single byte value
  out into a full word using the shufflebytes instruction.

- Clean up header include ordering to be more canonical, and remove
  spurious #undefs of function names.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The "inv" (invalidate) instruction is generally less safe than "finv"
(flush and invalidate), as it will drop dirty data from the cache.
It turns out we have almost no need for "inv" (other than for the
older 32-bit architecture in some limited cases), so convert to
"finv" where possible and delete the extra "inv" infrastructure.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

First, clean up active hardwalls in exit_thread().  This is a better
place than in arch_release_thread_info().

Second, mask out any non-online cpus from the cpumask after
validating any required semantics of the cpu set.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Commit 0998d063 (device-core: Ensure drvdata = NULL when no
driver is bound) removes the need to set driver data field to
NULL.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Need add cmpxchg64(), or will cause compiling issue.

Need define it as cmpxchg() only for 64-bit operation, since cmpxchg()
can support 8 bytes.

The related error (with allmodconfig):

  drivers/block/blockconsole.c: In function ‘bcon_advance_console_bytes’:
  drivers/block/blockconsole.c:164:2: error: implicit declaration of function ‘cmpxchg64’ [-Werror=implicit-function-declaration]
Signed-off-by: Chen Gang <gang.chen@asianux.com>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Pull slab update from Pekka Enberg:
 "Highlights:

  - Fix for boot-time problems on some architectures due to
    init_lock_keys() not respecting kmalloc_caches boundaries
    (Christoph Lameter)

  - CONFIG_SLUB_CPU_PARTIAL requested by RT folks (Joonsoo Kim)

  - Fix for excessive slab freelist draining (Wanpeng Li)

  - SLUB and SLOB cleanups and fixes (various people)"

I ended up editing the branch, and this avoids two commits at the end
that were immediately reverted, and I instead just applied the oneliner
fix in between myself.

* 'slab/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux
  slub: Check for page NULL before doing the node_match check
  mm/slab: Give s_next and s_stop slab-specific names
  slob: Check for NULL pointer before calling ctor()
  slub: Make cpu partial slab support configurable
  slab: add kmalloc() to kernel API documentation
  slab: fix init_lock_keys
  slob: use DIV_ROUND_UP where possible
  slub: do not put a slab to cpu partial list when cpu_partial is 0
  mm/slub: Use node_nr_slabs and node_nr_objs in get_slabinfo
  mm/slub: Drop unnecessary nr_partials
  mm/slab: Fix /proc/slabinfo unwriteable for slab
  mm/slab: Sharing s_next and s_stop between slab and slub
  mm/slab: Fix drain freelist excessively
  slob: Rework #ifdeffery in slab.h
  mm, slab: moved kmem_cache_alloc_node comment to correct place

In the -rt kernel (mrg), we hit the following dump:

BUG: unable to handle kernel NULL pointer dereference at           (null)
IP: [<ffffffff811573f1>] kmem_cache_alloc_node+0x51/0x180
PGD a2d39067 PUD b1641067 PMD 0
Oops: 0000 [#1] PREEMPT SMP
Modules linked in: sunrpc cpufreq_ondemand ipv6 tg3 joydev sg serio_raw pcspkr k8temp amd64_edac_mod edac_core i2c_piix4 e100 mii shpchp ext4 mbcache jbd2 sd_mod crc_t10dif sr_mod cdrom sata_svw ata_generic pata_acpi pata_serverworks radeon ttm drm_kms_helper drm hwmon i2c_algo_bit i2c_core dm_mirror dm_region_hash dm_log dm_mod
CPU 3
Pid: 20878, comm: hackbench Not tainted 3.6.11-rt25.14.el6rt.x86_64 #1 empty empty/Tyan Transport GT24-B3992
RIP: 0010:[<ffffffff811573f1>]  [<ffffffff811573f1>] kmem_cache_alloc_node+0x51/0x180
RSP: 0018:ffff8800a9b17d70  EFLAGS: 00010213
RAX: 0000000000000000 RBX: 0000000001200011 RCX: ffff8800a06d8000
RDX: 0000000004d92a03 RSI: 00000000000000d0 RDI: ffff88013b805500
RBP: ffff8800a9b17dc0 R08: ffff88023fd14d10 R09: ffffffff81041cbd
R10: 00007f4e3f06e9d0 R11: 0000000000000246 R12: ffff88013b805500
R13: ffff8801ff46af40 R14: 0000000000000001 R15: 0000000000000000
FS:  00007f4e3f06e700(0000) GS:ffff88023fd00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 00000000a2d3a000 CR4: 00000000000007e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process hackbench (pid: 20878, threadinfo ffff8800a9b16000, task ffff8800a06d8000)
Stack:
 ffff8800a9b17da0 ffffffff81202e08 ffff8800a9b17de0 000000d001200011
 0000000001200011 0000000001200011 0000000000000000 0000000000000000
 00007f4e3f06e9d0 0000000000000000 ffff8800a9b17e60 ffffffff81041cbd
Call Trace:
 [<ffffffff81202e08>] ? current_has_perm+0x68/0x80
 [<ffffffff81041cbd>] copy_process+0xdd/0x15b0
 [<ffffffff810a2125>] ? rt_up_read+0x25/0x30
 [<ffffffff8104369a>] do_fork+0x5a/0x360
 [<ffffffff8107c66b>] ? migrate_enable+0xeb/0x220
 [<ffffffff8100b068>] sys_clone+0x28/0x30
 [<ffffffff81527423>] stub_clone+0x13/0x20
 [<ffffffff81527152>] ? system_call_fastpath+0x16/0x1b
Code: 89 fc 89 75 cc 41 89 d6 4d 8b 04 24 65 4c 03 04 25 48 ae 00 00 49 8b 50 08 4d 8b 28 49 8b 40 10 4d 85 ed 74 12 41 83 fe ff 74 27 <48> 8b 00 48 c1 e8 3a 41 39 c6 74 1b 8b 75 cc 4c 89 c9 44 89 f2
RIP  [<ffffffff811573f1>] kmem_cache_alloc_node+0x51/0x180
 RSP <ffff8800a9b17d70>
CR2: 0000000000000000
---[ end trace 0000000000000002 ]---

Now, this uses SLUB pretty much unmodified, but as it is the -rt kernel
with CONFIG_PREEMPT_RT set, spinlocks are mutexes, although they do
disable migration. But the SLUB code is relatively lockless, and the
spin_locks there are raw_spin_locks (not converted to mutexes), thus I
believe this bug can happen in mainline without -rt features. The -rt
patch is just good at triggering mainline bugs ;-)

Anyway, looking at where this crashed, it seems that the page variable
can be NULL when passed to the node_match() function (which does not
check if it is NULL). When this happens we get the above panic.

As page is only used in slab_alloc() to check if the node matches, if
it's NULL I'm assuming that we can say it doesn't and call the
__slab_alloc() code. Is this a correct assumption?
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull more vfs stuff from Al Viro:
 "O_TMPFILE ABI changes, Oleg's fput() series, misc cleanups, including
  making simple_lookup() usable for filesystems with non-NULL s_d_op,
  which allows us to get rid of quite a bit of ugliness"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  sunrpc: now we can just set ->s_d_op
  cgroup: we can use simple_lookup() now
  efivarfs: we can use simple_lookup() now
  make simple_lookup() usable for filesystems that set ->s_d_op
  configfs: don't open-code d_alloc_name()
  __rpc_lookup_create_exclusive: pass string instead of qstr
  rpc_create_*_dir: don't bother with qstr
  llist: llist_add() can use llist_add_batch()
  llist: fix/simplify llist_add() and llist_add_batch()
  fput: turn "list_head delayed_fput_list" into llist_head
  fs/file_table.c:fput(): add comment
  Safer ABI for O_TMPFILE

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>