HardenedBSD/libexec/rtld-elf/alpha/rtld_start.S
John Polstra 630df077ab Solve the dynamic linker's problems with multithreaded programs once
and for all (I hope).  Packages such as wine, JDK, and linuxthreads
should no longer have any problems with re-entering the dynamic
linker.

This commit replaces the locking used in the dynamic linker with a
new spinlock-based reader/writer lock implementation.  Brian
Fundakowski Feldman <green> argued for this from the very beginning,
but it took me a long time to come around to his point of view.
Spinlocks are the only kinds of locks that work with all thread
packages.  But on uniprocessor systems they can be inefficient,
because while a contender for the lock is spinning the holder of the
lock cannot make any progress toward releasing it.  To alleviate
this disadvantage I have borrowed a trick from Sleepycat's Berkeley
DB implementation.  When spinning for a lock, the requester does a
nanosleep() call for 1 usec. each time around the loop.  This will
generally yield the CPU to other threads, allowing the lock holder
to finish its business and release the lock.  I chose 1 usec. as the
minimum sleep which would with reasonable certainty not be rounded
down to 0.

The formerly machine-independent file "lockdflt.c" has been moved
into the architecture-specific subdirectories by repository copy.
It now contains the machine-dependent spinlocking code.  For the
spinlocks I used the very nifty "simple, non-scalable reader-preference
lock" which I found at

  <http://www.cs.rochester.edu/u/scott/synchronization/pseudocode/rw.html>

on all CPUs except the 80386 (the specific CPU model, not the
architecture).  The 80386 CPU doesn't support the necessary "cmpxchg"
instruction, so on that CPU a simple exclusive test-and-set lock
is used instead.  80386 CPUs are detected at initialization time by
trying to execute "cmpxchg" and catching the resulting SIGILL
signal.

To reduce contention for the locks, I have revamped a couple of
key data structures, permitting all common operations to be done
under non-exclusive (reader) locking.  The only operations that
require exclusive locking now are the rare intrusive operations
such as dlopen() and dlclose().

The dllockinit() interface is now deprecated.  It still exists,
but only as a do-nothing stub.  I plan to remove it as soon as is
reasonably possible.  (From the very beginning it was clearly
labeled as experimental and subject to change.)  As far as I know,
only the linuxthreads port uses dllockinit().  This interface turned
out to have several problems.  As one example, when the dynamic
linker called a client-supplied locking function, that function
sometimes needed lazy binding, causing re-entry into the dynamic
linker and a big looping mess.  And in any case, it turned out to be
too burdensome to require threads packages to register themselves
with the dynamic linker.
2000-07-08 04:10:38 +00:00

219 lines
5.6 KiB
ArmAsm

/* $FreeBSD$ */
/* From: NetBSD: rtld_start.S,v 1.1 1996/12/16 20:38:09 cgd Exp */
/*
* Copyright 1996 Matt Thomas <matt@3am-software.com>
* Copyright 2000 John D. Polstra
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <machine/asm.h>
#include <machine/pal.h>
.extern _GLOBAL_OFFSET_TABLE_
LEAF(_rtld_start, 0) /* XXX */
.set noreorder
br pv, $33
$33: LDGP(pv)
/* save away the stack pointer */
lda s0, 0(sp) /* get argc from stack */
lda sp, -16(sp) /* space for arguments */
/* save ps_strings pointer */
mov a3, s1
/* Step 1 -- Figure out the displacement */
br t2, $34 /* get our PC */
$34: ldiq t3, $34 /* get where the linker thought we were */
subq t2, t3, t8 /* calculate the displacement */
/* Step 2 -- Find bounds of global offset table */
lda t5, _GLOBAL_OFFSET_TABLE_
addq t8, t5, t9 /* add the displacement */
lda t4, _DYNAMIC
addq t8, t4, t10 /* add the displacement */
/*
* Step 3 -- Every entry in the global offset table needs to
* modified for the displacement before any code will work.
*/
$35: ldq t1, 0(t9) /* load the value */
addq t8, t1, t1 /* add the displacement */
stq t1, 0(t9) /* save the new value */
lda t9, 8(t9) /* point to next entry */
cmpult t9, t10, t1 /* are we done? */
bne t1, $35 /* no, do more */
/*
* Ya! Things are far enough so we can do some dynamic linking!
*/
lda a0, 0(s0) /* initial sp */
lda a1, -16(s0) /* address for exit proc */
lda a2, -8(s0) /* address for obj_main */
CALL(_rtld) /* v0 = _rtld(sp, &exit_proc, &obj_main); */
ldq a1, -16(s0) /* our atexit function */
ldq a2, -8(s0) /* obj_main entry */
lda sp, 16(sp) /* readjust our stack */
mov s0, a0 /* stack pointer */
mov s1, a3 /* ps_strings pointer */
mov v0, t12
jsr ra, (v0), 0 /* (*_start)(sp, cleanup, obj); */
ldgp gp, 0(ra)
CALL(exit)
halt
END(_rtld_start)
.set noat
.globl _rtld_bind_start
.ent _rtld_bind_start
_rtld_bind_start:
lda sp, -168(sp)
.frame sp, 168, $26
/* Preserve all registers that C normally doesn't. */
stq $26, 0(sp)
stq $0, 8(sp)
stq $1, 16(sp)
stq $2, 24(sp)
stq $3, 32(sp)
stq $4, 40(sp)
stq $5, 48(sp)
stq $6, 56(sp)
stq $7, 64(sp)
stq $8, 72(sp)
stq $16, 80(sp)
stq $17, 88(sp)
stq $18, 96(sp)
stq $19, 104(sp)
stq $20, 112(sp)
stq $21, 120(sp)
stq $22, 128(sp)
stq $23, 136(sp)
stq $24, 144(sp)
stq $25, 152(sp)
stq $29, 160(sp)
.mask 0x27ff01ff, -168
/* Set up our $gp */
br gp, $100
$100: ldgp gp, 0(gp)
.prologue 1
/* Set up the arguments for _rtld_bind. */
subq at_reg, t12, a1 /* calculate reloc offset */
ldq a0, 8(t12) /* object structure */
subq a1, 20, a1 /* = (at - t11 - 20) / 12 * 24 */
addq a1, a1, a1
CALL(_rtld_bind)
/* Move the destination address into position. */
mov $0, $27
/* Restore program registers. */
ldq $26, 0(sp)
ldq $0, 8(sp)
ldq $1, 16(sp)
ldq $2, 24(sp)
ldq $3, 32(sp)
ldq $4, 40(sp)
ldq $5, 48(sp)
ldq $6, 56(sp)
ldq $7, 64(sp)
ldq $8, 72(sp)
ldq $16, 80(sp)
ldq $17, 88(sp)
ldq $18, 96(sp)
ldq $19, 104(sp)
ldq $20, 112(sp)
ldq $21, 120(sp)
ldq $22, 128(sp)
ldq $23, 136(sp)
ldq $24, 144(sp)
ldq $25, 152(sp)
ldq $29, 160(sp)
/* Flush the Icache after having modified the .plt code. */
imb
/* Clean up and turn control to the destination */
lda sp, 168(sp)
jmp $31, ($27)
.end _rtld_bind_start
/*
* int cmp0_and_store_int(volatile int *p, int newval);
*
* If an int holds 0, store newval into it; else do nothing. Returns
* the previous value.
*/
LEAF(cmp0_and_store_int, 2)
1: mov a1, t0
ldl_l v0, 0(a0)
bne v0, 3f
stl_c t0, 0(a0)
beq t0, 2f
mb
RET
2: br 1b
3: RET
END(cmp0_and_store_int)
LEAF(atomic_add_int, 2)
0: ldl_l t0, 0(a0)
addq t0, a1, t0
stl_c t0, 0(a0)
beq t0, 1f
mb
RET
1: br 0b
END(atomic_add_int)
/* Atomically increment an int. */
LEAF(atomic_incr_int, 1)
0: ldl_l t0, 0(a0)
addq t0, 1, t0
stl_c t0, 0(a0)
beq t0, 1f
mb
RET
1: br 0b
END(atomic_incr_int)
/* Atomically decrement an int. */
LEAF(atomic_decr_int, 1)
0: ldl_l t0, 0(a0)
subq t0, 1, t0
stl_c t0, 0(a0)
beq t0, 1f
mb
RET
1: br 0b
END(atomic_decr_int)