Add the TIF_NOHZ flag appropriately.

Add call to user_exit() on entry to do_work_pending() and on entry
to syscalls via do_syscall_trace_enter(), and also the top of
do_syscall_trace_exit() just because it's done in x86.

Add call to user_enter() at the bottom of do_work_pending() once we
have no more work to do before returning to userspace.

Wrap all the trap code in exception_enter() / exception_exit().

Signed-off-by: Chris Metcalf <cmetcalf@ezchip.com>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>

The r1 value is set based on the r0 value as we return to user space.
So tracing tools won't automatically see the right value.  Fix this by
generating the correct r1 value in do_syscall_trace_exit() rather
than trying to tamper with the hot path in syscall return.

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>

Currently the syscall number is passed, but it should be the return
value, which is kept in r0.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com> [using a raw 0 value]

This patch adds support for the TIF_SYSCALL_TRACEPOINT on the tile
architecture. Basically, it calls the appropriate tracepoints on syscall
entry and exit.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Call tracehook functions for syscall tracing.

The check for TIF_SYSCALL_TRACE was removed, because the same check is
done right before in the assembly file.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com> [with ptrace.h fixup]

This patch changes arch_ptrace on tile so that it uses user_regset
to implement the PTRACE_GETREGS and PTRACE_SETREGS operations.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This is a basic implementation of user_regset for the tile
architecture. It reuses the basic blocks that were already there.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

Use the newer idioms for setting PTRACE_O_xxx and PT_TRACE_xxx flags.
Only set/clear tile-specific flags if the generic routine returns
success, since otherwise we want to avoid setting any flags at all.
Atomically update the ptrace flags with the new values.  Eliminate
the PT_TRACE_MASK_TILE bitmask and just shift PTRACE_O_MASK_TILE.
Add a BUILD_BUG_ON to avoid overlapping with generic bits.

Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

This flag is set for ptrace GETREGS or PEEKUSER for processes
that are COMPAT, i.e. 32-bit.  This allows things like strace
to easily discover what personality to use, for example.

Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>