blob: 9b989eb2dac1bfae515afbc0fb926cbca997dab9 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
|
# $NetBSD: options.mk,v 1.7 2008/05/31 20:43:29 tnn Exp $
PKG_OPTIONS_VAR= PKG_OPTIONS.MesaLib
PKG_SUPPORTED_OPTIONS= mesa-execmem-mmap
# Assembler code build configurations
.if (${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "x86_64") && \
${OPSYS} != "SunOS" && ${OPSYS} != "Darwin"
PKG_SUPPORTED_OPTIONS+= ${MACHINE_ARCH}
PKG_SUGGESTED_OPTIONS+= ${MACHINE_ARCH}
.endif
###
### XXX Prior to this patch, Mesa only allocated executable memory
### with mmap(2) and PROT_EXEC on linux. On e.g. OpenBSD and NetBSD,
### pages which absolutely required PROT_EXEC were allocated using
### malloc(3). You may wonder why it only worked on Linux. Answer:
### the code was excluded with '#ifdef __linux__'! "Porting" this
### [to BSD platforms] involved adding three lines of [preprocessor]
### code (see patch-ap if you're curious).
###
### Please do note hack (NetBSD-only for now) I added to wire down
### this memory by setting the MESA_EXECMEM_MMAP environment variable;
### you probably don't want it paged out. MAP_WIRED is probably
### overkill--I will revisit this at a later date.
###
### You may also define MESA_EXECMEM_HEAP_SIZE in your build
### environment to adjust the size of the anonymous executable
### memory pool (in bytes). The default is 10MB (huge?).
###
### Here's the bottom line: IF YOUR PLATFORM DOES NOT HAVE AN
### EXECUTABLE HEAP, YOU RUN THE RISK OF MANY SEGMENTATION FAULTS
### UNLESS THIS OPTION IS ENABLED. If enabling this option results in
### build failure, please do file a PR. Thanks to Owain Ainsworth
### <oga@openbsd.org> for discovering this.
###
### --bjs (04/23/08)
###
.if (!empty(OPSYS:M*BSD) || ${OPSYS} == "DragonFly" || ${OPSYS} == "Linux")
PKG_SUGGESTED_OPTIONS= mesa-execmem-mmap
.endif
###
### XXX There are [probably] others, but let's not get crazy just yet.
### This will take a while to test for the myriad platforms we
### support.
###
.if (${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "x86_64" || \
${MACHINE_ARCH} == "powerpc" || ${MACHINE_ARCH} == "sparc" || \
${MACHINE_ARCH} == "sparc64") && \
(${OPSYS} == "NetBSD" || ${OPSYS} == "FreeBSD" || ${OPSYS} == "OpenBSD" || \
${OPSYS} == "DragonFly")
PKG_SUPPORTED_OPTIONS+= dri
.endif
###
### XXX OpenGL still works fine with the software fallback. As of now,
### I think this is a good way to see which bugs surface before the
### next release branch. Upgrading the X server to the 1.4 branch
### is advised given that it's glx/glcore modules are built from
### Mesa 6.5.3 (a development release).
###
.if !empty(MACHINE_PLATFORM:MNetBSD-[4-9]*-*86*)
PKG_SUGGESTED_OPTIONS+= dri
.endif
.include "../../mk/bsd.options.mk"
###
### XXX Yes, this is a bit overly verbose; with Mesa, that can't hurt much.
###
.if !empty(PKG_OPTIONS:Mi386) || !empty(PKG_OPTIONS:Mx86_64)
BUILD_TARGET_SUFFIX= -${MACHINE_ARCH}
.else
BUILD_TARGET_SUFFIX= # empty
.endif
.if !empty(PKG_OPTIONS:Mdri)
BUILD_TARGET= pkgsrc-dri${BUILD_TARGET_SUFFIX}
PLIST.dri= # empty
. include "../../graphics/MesaLib/dri.mk"
.else
BUILD_TARGET= pkgsrc${BUILD_TARGET_SUFFIX}
PLIST.nodri= # empty
###
### XXX building libOSMesa breaks with -j, and GNU make has no .WAIT
###
MAKE_JOBS_SAFE= no
.endif
###
### XXX This is the default heap size. Would there be an occasion to
### change it? It seems large, so we should investigate further.
###
MESA_EXECMEM_HEAPSIZE?= 10485760
.if !empty(PKG_OPTIONS:Mmesa-execmem-mmap)
CFLAGS+= -DMESA_EXECMEM_MMAP
CFLAGS+= -DEXEC_HEAP_SIZE=${MESA_EXECMEM_HEAPSIZE:M[0-9]*:Q}
.endif
|
