Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
148,029 Commits 606 Branches 0 Tags 3.1 GiB
5fe28dc7e5
Commit Graph

424 Commits

Author SHA1 Message Date
ross
520b9c1f59 * remove .ALLSRC, which doesn't work with the new dependency tree 
replaced mostly with .IMPSRC, so add .c wrappers for variant .o names

* any machine with a shared crt0.c that also uses common_elf (I didn't see
  any) will now need a reachover wrapper in .CURDIR a la crtbeginS.c

this should fix the build breakage
 2003-03-19 04:31:41 +00:00
christos
e04b01c9ff Add appropriate SRCS lines so that we actually create dependencies. I've 
been running with a crt0.o from last year!
 2003-03-18 19:11:57 +00:00
scw
699eeb228c The temporary asm trampoline code is no longer needed. 2002-12-06 17:06:06 +00:00
scw
098d74e947 Fix the alignment of the MD_SECTION_PROLOGUE. 2002-12-06 17:05:14 +00:00
thorpej
f343e6c086 If EXTERNAL_TOOLCHAIN is defined, don't build crtbegin/crtend. 2002-11-26 19:27:09 +00:00
thorpej
df201a40b9 Correct a comment. 2002-11-23 17:23:51 +00:00
thorpej
5df3ea5331 Oops, the _init() and _fini() entry points need to be in the .init 
and .fini sections, respectively (same as the {init,fini}_fallthru
stubs).

This is needed on PowerPC, where otherwise the linker would be unhappy
with the relocation records for the fallthru stubs when building a
shared library.
 2002-11-23 17:21:22 +00:00
thorpej
bc4956ce1c Fix the _init and _fini function prologues for ABICALLS. Make sure 
the functions are aligned to a 32-bit boundary, otherwise some pretty
colorful lossage can result.
 2002-11-23 01:44:13 +00:00
thorpej
6a088e4e31 Fix typo. 2002-11-22 16:47:20 +00:00
thorpej
6dba3a7439 Split crtbegin/crtend into crti/crtbegin/crtend/crtn. NetBSD-specific 
things like the .note.netbsd.ident section are provided by crti/crtn.
crti/crtn also provide the _init() and _fini() routines.

crtbegin/crtend now only provide support for ctors/dtors.  This paves
the way to using the "crtstuff" provided with GCC (when we upgrade to
GCC 3.3), which provides, among other things, much better C++/Java
exception handling.
 2002-11-22 06:44:56 +00:00
itohy
cbb1f0fb0f MKDYNAMICROOT for a.out 2002-11-12 14:33:48 +00:00
itohy
5b3e8cb3f2 Do not print \0 at the end of error message. 2002-11-12 14:26:10 +00:00
thorpej
303696ca38 Fix signed/unsigned comparison warnings. 2002-11-11 00:44:43 +00:00
grant
291ac7e9c4 don't include machine/asm.h (make this build again). 2002-09-19 01:24:43 +00:00
lukem
ebb6fc9eb8 Use ${NETBSDSRCDIR}/some/path instead of ${.CURDIR}/../../some/path (etc). 
(Reduces make output by ~ 20%)
 2002-08-19 09:41:27 +00:00
matt
b79865a0af To make things clearer, use register prefixes in asm() code. 2002-07-29 21:54:35 +00:00
scw
c7b95fdc37 Use an assembler trampoline to call the real start entrypoint, at least 
until we switch to the native toolchain.
 2002-07-10 13:13:47 +00:00
scw
59474a8c82 NetBSD, meet the SH-5 cpu. 
SH-5, meet NetBSD.

Let's hope this is the start of a long and fruitful relationship. :-)

This code, funded by Wasabi Systems, adds initial support for the
Hitachi SuperH(tm) SH-5 cpu architecture to NetBSD.

At the present time, NetBSD/evbsh5 only runs on a SH-5 core simulator
which has no simulated devices other than a simple console. However, it
is good enough to get to the "root device: " prompt.

Device driver support for Real SH-5 Hardware is in place, particularly for
supporting the up-coming Cayman evaluation board, and should be quite
easy to get running when the hardware is available.

There is no in-tree toolchain for this port at this time. Gcc-current has
rudimentary SH-5 support but it is known to be buggy. A working toolchain
was obtained from SuperH to facilitate this port. Gcc-current will be
fixed in due course.

The SH-5 architecture is fully 64-bit capable, although NetBSD/evbsh5 has
currently only been tested in 32-bit mode. It is bi-endian, via a boot-
time option and it also has an "SHcompact" mode in which it will execute
SH-[34] user-land instructions.

For more information on the SH-5, see www.superh.com. Suffice to say it
is *not* just another respin of the SH-[34].
 2002-07-05 13:31:28 +00:00
fredette
43c563a3dd A new crt0.c for hppa, with support for dynamic linking. 2002-07-01 15:56:41 +00:00
matt
0ec15f80cc Allow this to compile under gcc-current. No multiline strings for asm. 2002-06-09 06:18:34 +00:00
thorpej
8bbbfff0a1 Remove __LEADING_UNDERSCORE case. 2002-06-08 17:27:25 +00:00
fredette
02ac1ae2f0 Added hppa support (some of it incomplete) to lib/csu, lib/libc, 
and lib/libkvm.
 2002-06-06 20:31:19 +00:00
fvdl
e4b98ed506 Add init/fini handling. 2002-06-06 19:18:22 +00:00
fvdl
e8d64f072a -x is too heavy-handed for removing symbols; newer versions of GNU ld 
will completely delete empty local symbols with it. Use -X instead
for crt{begin,end}*.o, to keep init_fallthru hacks.
 2002-06-06 19:17:01 +00:00
thorpej
7f350cba6c Don't use multi-line string literals. 2002-06-05 02:51:03 +00:00
wiz
da0b574ae0 Comment out token after #endif. 2002-05-16 19:30:41 +00:00
matt
858f46fb43 Define _SDA(2)_BASE_ as [] which is more correct. 2002-05-09 20:32:59 +00:00
thorpej
6249662a70 Deal better with platforms that don't yet support shared libraries: 
* Only add -fPIC if MKPIC != no.
* Only build crtbeginS and crtendS if MKPIC != no.
 2002-05-05 00:45:31 +00:00
thorpej
179139677a All ELF toolchains are of the "new toolchain" variety now, so no need 
to conditionalize use of .hidden symbols anymore.
 2002-05-05 00:34:00 +00:00
wiz
15d53d85be ${MACHINE_ARCH} != "arm32" is always true now, so remove it from an if. 2002-05-02 16:42:53 +00:00
matt
e18dedbd2f switch from addis R,0,... to the simplified mnemonic of lis which is 
typically used for this.
 2002-04-26 23:28:53 +00:00
kleink
3f1a2352b7 Use __weak_extern() for _SDA_BASE_ and _SDA2_BASE_ declarations. 2002-04-17 17:36:13 +00:00
matt
036d79bb12 Make sure to initialize GPR13/GPR2 appropriately for the Small Data Area 
pointer as described in the SVR4 PPC ABI.  Though GPR2 is defined by the
SVR4 PPC ABI as system-reserved, the EABI (Embedded API) changes its use
to be the Small Data Area 2 register (const small data).  Since powerpc
under NetBSD doesn't use R2 in the kernel (yet), support it as well.
 2002-04-17 13:32:42 +00:00
skrll
2d0a923dcd Correct some comments. 2002-04-08 22:17:37 +00:00
jmc
c03ef26393 Don't make ___start static as gcc can't tell it's referenced and complains. 2002-04-04 08:22:11 +00:00
eeh
41065b740e Don't need start' in crt0.o any more. ld expects _start' as the entry 
symbol, while gcc likes `__start'.
 2002-03-29 18:11:55 +00:00
matt
0ce5ca145f Change to use a register prefix. 2002-02-24 01:06:18 +00:00
matthias
21fb786623 Declare the global register variable kps_strings before including anything 
as this needs to be done before any function is defined (and there seems
to be a function definition in one of the include files).
 2002-02-22 13:46:30 +00:00
mrg
829b76292e fix multiline string literal. from gcc3 2002-02-11 06:08:23 +00:00
thorpej
e4217fbfe4 Pass the rtld object and ps_strings to _start() properly. From Marcus. 2002-02-03 23:57:19 +00:00
thorpej
ec206ac99b Multi-line string literals are not in ISO-C. 2002-01-23 03:55:33 +00:00
drochner
e757f639d2 don't include <machine/asm.h> to avoid conflicts with definitions 
in <sys/cdefs.h>
 2002-01-14 12:06:49 +00:00
thorpej
fd9112a780 Build and install crtfm.o (sets the Alpha FP_C in a way that GCC's 
-ffast-math on Alpha expects).
 2002-01-14 01:31:58 +00:00
thorpej
24ca47c4d5 * Use the new ALPHA_SET_FP_C sysarch call. 
* Provide an internal sysarch entry point, so that this works
  with statically-linked programs.
 2002-01-14 01:31:00 +00:00
thorpej
655925395a Support for the GCC "-ffast-math" option on the Alpha. 2002-01-13 20:05:42 +00:00
thorpej
caaec27cfe * Use += to assign OBJS so that machine-dependent Makefiles can 
add their own objects to the list.
* Garbage-collect an unneeded CLEANFILES assignment.
 2002-01-13 19:59:26 +00:00
thorpej
d605da06f0 %progbits, not @progbits (@ is a comment character in ARM assembly). 2002-01-01 01:36:21 +00:00
thorpej
7050ac9aa9 Multi-line string literals are not supported by ANSI C. 2002-01-01 01:31:06 +00:00
thorpej
5ced70565e Set ELFSIZE=64. 2002-01-01 01:14:22 +00:00
thorpej
d8abc59932 Add support for invoking GNU C++ local destructors in shared objects 
via the __cxa_finalize() mechanism.

XXX Only enabled for new-toolchain platforms, as it requires the assembler
to support the .hidden pseudo-op.
 2001-12-31 00:40:11 +00:00
thorpej
a50163af65 Build separate crt{begin,end}.o and crt{begin,end}S.o files, rather 
than linking them at install time.  Forthcoming changes require different
behavior for static vs. shared object versions of these files.
 2001-12-31 00:11:13 +00:00
thorpej
c41160abff Add support for registering Java classes emitted by the Java compiler 
(e.g. the Java compiler in GCC 3.x) into the ".jcr" section.
 2001-12-30 23:45:00 +00:00
thorpej
6027dc5536 Make sure the Dwarf2 EH frame info is 4-byte aligned. 2001-12-30 23:26:20 +00:00
thorpej
5b81869561 Don't use multi-line string literals; they're not supported by ANSI C. 2001-12-30 23:23:38 +00:00
bjh21
7723559395 Quiet GCC warnings. 2001-11-08 22:23:59 +00:00
lukem
65469e54b3 pull in ../../Makefile.inc 2001-11-05 15:11:40 +00:00
skrll
6d6c931c5a This is an a.out thing and doesn't belong here. 2001-10-26 16:05:25 +00:00
scw
f768a5c42b Reduce the alignment constraint from 4 to 2. This fixes the problem where 
gas(2) pads the text segment with 0x0 instead of NOP. Since m68k instructions
only need 16-bit alignment, .align 2 works just fine.
 2001-10-13 09:21:50 +00:00
kristerw
031a62292d Run the entries in __CTOR_LIST__ and __DTOR_LIST__ in the order gcc expect. 2001-10-11 18:02:28 +00:00
matt
eef4b63246 Make __start global. 2001-09-08 18:59:20 +00:00
bjh21
751e6fa590 Ensure that the stack is aligned to an 8-byte boundary before calling C code. 
(ATPCS B-01)
 2001-08-17 00:14:38 +00:00
simonb
9a68629ca6 Enlarge the stack frame to 16 bytes, and store the return address 8 
bytes in instead of at the start, to leave room for a .cpload to store
the gp at offset 0 in the frame.  Allow 8 bytes for each (for mips64
one day...).

.cpload overwrite problems noted by Michael Hitch.
 2001-08-16 03:44:46 +00:00
matt
ee6a16ad18 Add a secondary label called __start (makes gcc 2.95.3 happier) and rename 
the internal __start to ___start.   (Since -e is not supplied by default,
this has no effect on a native build.)
 2001-08-05 00:01:27 +00:00
thorpej
102df4197d Enable DWARF2 EH support. Confirmed working w/ GCC 2.95.3 on sparc 
and i386.
 2001-08-03 14:26:13 +00:00
thorpej
a6855eea24 Add support for DWARF2 frame unwinding/exception handling. This 
should be sufficient for GCC 2.95.3, and probably GCC 3.0 as well.

Not yet enabled by default; further testing required.
 2001-08-03 05:54:43 +00:00
wiz
e11d04474c issueing -> issuing 2001-07-26 22:55:12 +00:00
tsubai
1278735b51 .init/.fini support. 2001-07-23 20:43:31 +00:00
bjh21
71d647ec7a The common arm_elf code seems to work on arm26, so use that. 2001-07-20 13:01:13 +00:00
simonb
19211890ff Enable dot_init.h on mips, fixing .init & .fini processing. 2001-07-18 04:43:48 +00:00
mrg
6bea810c23 enable dot_init.h on both sparc & sparc64, fixing .init & .fini processing 
for both platforms.  they share dot_init.h with a single #ifdef #define
difference...
 2001-07-17 18:25:11 +00:00
simonb
1bc139fd41 Add but not enable correct .init/.fini handling. 2001-07-17 15:20:23 +00:00
mrg
f4bd34ec5f add but not enable correct .init/.fini handling. 2001-07-17 13:32:35 +00:00
mrg
ff68bb147c minor KNF 2001-07-17 13:28:05 +00:00
tsubai
4a0b75d5d2 Add the new .init/.fini support. 2001-07-17 12:16:43 +00:00
matt
119c72fb96 Add new init/fini to VAX ELF. 2001-07-17 06:45:46 +00:00
matt
ae9f64ff3b Turn-on new init/fini processing. 2001-07-17 06:39:14 +00:00
thorpej
4abcdae85e Enable the new (working) .init/.fini magic. 2001-07-17 03:56:48 +00:00
matt
126898a2cd Initial CSU for ARM with shared ELF. 2001-07-15 17:29:00 +00:00
fvdl
a84b13f43f CSU for x86-64 port. 2001-06-19 00:49:06 +00:00
mrg
f59b9e737f set FILESDIR to "${LIBDIR}", not hard coded /usr/lib. 2001-06-07 15:22:49 +00:00
fredette
e80c95ddb2 Don't use an EA with a scaled index, it isn't supported on the 68000. 2001-05-17 21:21:07 +00:00
fredette
872a39428c Don't compile with -fPIC when MACHINE_ARCH is m68000. 2001-05-17 21:17:02 +00:00
kleink
7bd28edec7 Add, but do not yet enable, m68k code for .init and .fini section fallthroughs. 2001-05-12 00:17:55 +00:00
ross
2d62241cb9 Reorganization of crtbegin/crtend. 
o  Primary mod: alpha and i386 code for .init and .fini section fallthroughs
o  Actual fallthru is not yet enabled, see csu/{alpha,i386_elf}/Makefile
o  When enabled, should fix bin/12865 & should fix at least one WINE bug,
   see http://www.winehq.com/hypermail/wine-devel/2001/05/0021.html
o  Also, removed the function pointer hack put in to prevent gcc from
   incorrectly attempting to do short calls over section boundaries,
   on some architectures, rather, simply simply move ctors() and dtors()
   into .init and .fini.
Other notes.
o  Our crtbegin.c and crtend.c were completely in C, despite
   the fact that .init and .fini section fallthroughs require machine code.
   These modules implement the gnu ctors/dtors mechanism directly, eliminating
   in part the need for the elf .init/.fini section fallthrough. However,
   if we ever want to use the gnu-provided constructor support, we need to
   do the fallthrough, at least one program (WINE) apparently uses the
   feature independently, and other compilers may depend on the mechanism.
o  It is amusing that the previous NetBSD _init() and _fini() were put
   into the .init and .fini sections, because nothing is gained over
   plain .text if they don't fall through. Cargo cult!
o  Other elf ports will need to copy the alpha and i386 example mods.
 2001-05-11 22:44:14 +00:00
bjh21
0113cca535 Our ARM ELF ld(1) currently has problems with PIC relocations, so don't 
generate them.  I suspect this is safe until we get PIC in general working.
 2001-04-28 17:32:33 +00:00
msaitoh
581638d47c check __LEADING_UNDERSCORE 2001-02-25 15:51:24 +00:00
eeh
6a9224a06d Remove old compatibility hack. Should no longer be needed. 2001-02-21 00:47:21 +00:00
simonb
71c90eeec6 ANSIfy and merge some stylistic changes from arm26/crt0.c. 
No functional changes.
 2001-02-18 23:03:34 +00:00
simonb
5df49e36f3 Fix typo. 2001-02-18 23:01:39 +00:00
bjh21
015575c577 Use _C_LABEL rather than _C_FUNC. 2001-02-03 16:25:26 +00:00
enami
0ebf8f7f72 - Fix typo 
- Wrap long line
 2001-01-10 13:20:17 +00:00
itojun
99b0df5333 use MACIHNE_CPU to determine architecture-dependent source code directory. 
is use of exists(foo) safe enough?
TODO: many other places to fix
 2001-01-08 08:29:59 +00:00
itojun
264cf8dcf3 sh3 split 2001-01-04 08:01:49 +00:00
mycroft
94581370f5 The ELF ABI declares that argc must fill an `argument slot'. Make it so. 
For __sparc_v9__ only, do a hack to make executables work temporarily before
and after the kernel is switched.
 2000-07-26 15:40:07 +00:00
mycroft
209e309de1 Always dereference argc as an int. 2000-07-26 02:11:44 +00:00
mycroft
0ef3f52396 Make this work before and after the kernel is switched to DTRT with argc. 
This is temporary, only to make the flag day less painful.
 2000-07-25 21:42:55 +00:00
eeh
f310f1fd32 Handle dynamic binaries too. 2000-07-18 22:35:45 +00:00
matt
d71ca2d66e __syscall takes a quad_t as its first argument, not an int. 
_strrchr takes an int (due to promotion) as its third arg.
don't print the null byte at the end of string.
error messages should end with a linefeed.
 2000-07-03 23:01:01 +00:00
matt
06f872af34 Use the ___start etc. 2000-07-03 03:56:56 +00:00
christos
3d65200a80 remove initialization of mainprog_obj from here; it is too late for .ini 
sections that want to use dlopen(3).
 2000-06-16 20:52:49 +00:00