Mon Jan 6 19:13:07 2014 UTC ()
tickets 1009 and 1010


(bouyer)
diff -r1.1.2.11 -r1.1.2.12 src/doc/CHANGES-6.0.4

cvs diff -r1.1.2.11 -r1.1.2.12 src/doc/Attic/CHANGES-6.0.4 (switch to unified diff)

--- src/doc/Attic/CHANGES-6.0.4 2013/12/17 22:40:04 1.1.2.11
+++ src/doc/Attic/CHANGES-6.0.4 2014/01/06 19:13:07 1.1.2.12
@@ -1,133 +1,147 @@ @@ -1,133 +1,147 @@
1# $NetBSD: CHANGES-6.0.4,v 1.1.2.11 2013/12/17 22:40:04 riz Exp $ 1# $NetBSD: CHANGES-6.0.4,v 1.1.2.12 2014/01/06 19:13:07 bouyer Exp $
2 2
3A complete list of changes from the NetBSD 6.0.3 release to the NetBSD 6.0.4 3A complete list of changes from the NetBSD 6.0.3 release to the NetBSD 6.0.4
4release: 4release:
5 5
6doc/README.files patched by hand 6doc/README.files patched by hand
7gnu/usr.bin/groff/tmac/mdoc.local patched by hand 7gnu/usr.bin/groff/tmac/mdoc.local patched by hand
8sys/sys/param.h patched by hand 8sys/sys/param.h patched by hand
9 9
10 Welcome to 6.0.3_PATCH. 10 Welcome to 6.0.3_PATCH.
11 [jdc] 11 [jdc]
12 12
13xsrc/external/mit/xorg-server/dist/dix/dixfonts.c 1.2 13xsrc/external/mit/xorg-server/dist/dix/dixfonts.c 1.2
14xsrc/xfree/xc/programs/Xserver/dix/dixfonts.c 1.4 14xsrc/xfree/xc/programs/Xserver/dix/dixfonts.c 1.4
15 15
16 Fix CVE-2013-4396 using a patch from Alan Coopersmith: 16 Fix CVE-2013-4396 using a patch from Alan Coopersmith:
17 Save a pointer to the passed in closure structure before copying it 17 Save a pointer to the passed in closure structure before copying it
18 and overwriting the *c pointer to point to our copy instead of the 18 and overwriting the *c pointer to point to our copy instead of the
19 original. If we hit an error, once we free(c), reset c to point to 19 original. If we hit an error, once we free(c), reset c to point to
20 the original structure before jumping to the cleanup code that 20 the original structure before jumping to the cleanup code that
21 references *c. 21 references *c.
22 [spz, ticket #966] 22 [spz, ticket #966]
23 23
24sys/arch/x86/pci/pci_machdep.c 1.61 via patch 24sys/arch/x86/pci/pci_machdep.c 1.61 via patch
25 25
26 Force PCI mode 1 when running under QEMU, to work around 26 Force PCI mode 1 when running under QEMU, to work around
27 QEMU bug 897771. 27 QEMU bug 897771.
28 This should also make it possible to boot NetBSD under versions of KVM 28 This should also make it possible to boot NetBSD under versions of KVM
29 that have inherited said QEMU bug. Fixes PR kern/45671. 29 that have inherited said QEMU bug. Fixes PR kern/45671.
30 [gson, ticket #963] 30 [gson, ticket #963]
31 31
32sys/netinet/tcp_usrreq.c 1.168 32sys/netinet/tcp_usrreq.c 1.168
33 33
34 PR/48098: Brian Marcotte: Avoid kernel assertion for embryonic sockets 34 PR/48098: Brian Marcotte: Avoid kernel assertion for embryonic sockets
35 that don't have credentials yet. 35 that don't have credentials yet.
36 [spz, ticket #967] 36 [spz, ticket #967]
37 37
38external/mit/xorg/server/drivers/xf86-video-intel/Makefile 1.11 38external/mit/xorg/server/drivers/xf86-video-intel/Makefile 1.11
39 39
40 Add missing i810_dri.c file to SRCS. 40 Add missing i810_dri.c file to SRCS.
41 PR xsrc/48315. 41 PR xsrc/48315.
42 [martin, ticket #971] 42 [martin, ticket #971]
43 43
44distrib/utils/sysinst/mbr.c 1.92 44distrib/utils/sysinst/mbr.c 1.92
45 45
46 Add missing braces, which caused the offset of MBR partition 0 to be 46 Add missing braces, which caused the offset of MBR partition 0 to be
47 unintentionally set to 2048 even on small (<=128GB) disks. 47 unintentionally set to 2048 even on small (<=128GB) disks.
48 PR/48304. 48 PR/48304.
49 [tsutsui, ticket #972] 49 [tsutsui, ticket #972]
50 50
51sys/arch/xen/xen/xbdback_xenbus.c 1.58 51sys/arch/xen/xen/xbdback_xenbus.c 1.58
52 52
53 Fix a dom0 panic, or crash with a hypervisor panic, when creating 53 Fix a dom0 panic, or crash with a hypervisor panic, when creating
54 some domUs (domUs where the frontend driver doesn't provide a 54 some domUs (domUs where the frontend driver doesn't provide a
55 "protocol" entry in the xenstore). 55 "protocol" entry in the xenstore).
56 [bouyer, ticket #974] 56 [bouyer, ticket #974]
57 57
58libexec/ld.elf_so/tls.c 1.9 58libexec/ld.elf_so/tls.c 1.9
59tests/lib/libc/tls/dso/h_tls_dlopen.c 1.5 59tests/lib/libc/tls/dso/h_tls_dlopen.c 1.5
60 60
61 Only initialise TLS space from the PT_TLS segment, if the size is 61 Only initialise TLS space from the PT_TLS segment, if the size is
62 positive and the offset has been computed. Fixes PR lib/48324. 62 positive and the offset has been computed. Fixes PR lib/48324.
63 [joerg, ticket #976] 63 [joerg, ticket #976]
64 64
65usr.sbin/npf/npfctl/npf_ncgen.c patch 65usr.sbin/npf/npfctl/npf_ncgen.c patch
66sys/net/npf/npf_instr.c patch 66sys/net/npf/npf_instr.c patch
67 67
68 fix the byteorder for port range comparison 68 fix the byteorder for port range comparison
69 [rmind, ticket #986] 69 [rmind, ticket #986]
70 70
71sys/kern/uipc_socket.c 1.220 71sys/kern/uipc_socket.c 1.220
72 72
73 PR/48098: Brian Marcotte: 73 PR/48098: Brian Marcotte:
74 panic: kernel diagnostic assertion "cred != NULL": 74 panic: kernel diagnostic assertion "cred != NULL":
75 Fix from Michael van Elst, tcpdrop crashes kernel on 75 Fix from Michael van Elst, tcpdrop crashes kernel on
76 ebryonic connections. 76 ebryonic connections.
77 [spz, ticket #988] 77 [spz, ticket #988]
78 78
79sys/uvm/uvm_km.c 1.125 79sys/uvm/uvm_km.c 1.125
80 80
81 uvm_km_kmem_alloc: don't hardcode kmem_va_arena. This could lead to 81 uvm_km_kmem_alloc: don't hardcode kmem_va_arena. This could lead to
82 freeing to the wrong vmem arena in case of failed page allocations. 82 freeing to the wrong vmem arena in case of failed page allocations.
83 [para, ticket #989] 83 [para, ticket #989]
84 84
85sys/arch/sparc64/sparc64/locore.s 1.350 85sys/arch/sparc64/sparc64/locore.s 1.350
86 86
87 Increase an interrupt depth only in the case of hardware interrupts, 87 Increase an interrupt depth only in the case of hardware interrupts,
88 and remove the ci_idepth trick in softint_fastintr. 88 and remove the ci_idepth trick in softint_fastintr.
89 Fixes the following diagnostic panic reported in port-sparc64. 89 Fixes the following diagnostic panic reported in port-sparc64.
90 panic: kernel diagnostic assertion "!cpu_intr_p()" failed: file 90 panic: kernel diagnostic assertion "!cpu_intr_p()" failed: file
91 "../../../../kern/subr_xcall.c", line 351 91 "../../../../kern/subr_xcall.c", line 351
92 [nakayama, ticket #994] 92 [nakayama, ticket #994]
93 93
94sys/kern/uipc_syscalls.c 1.163 94sys/kern/uipc_syscalls.c 1.163
95 95
96 PR/47591: Michael Plass: If the unix socket is closed before accept, 96 PR/47591: Michael Plass: If the unix socket is closed before accept,
97 unp->unp_conn will be NULL in PRU_ACCEPT, as called from 97 unp->unp_conn will be NULL in PRU_ACCEPT, as called from
98 sys_accept->so_accept. This will cause the usrreq to return with 98 sys_accept->so_accept. This will cause the usrreq to return with
99 no error, leaving the mbuf gotten from m_get() with an uninitialized 99 no error, leaving the mbuf gotten from m_get() with an uninitialized
100 length, containing junk from a previous call. Initialize m_len to 100 length, containing junk from a previous call. Initialize m_len to
101 be 0 to handle this case.  101 be 0 to handle this case.
102 [spz, ticket #996] 102 [spz, ticket #996]
103 103
104sys/netinet6/nd6.c 1.146 104sys/netinet6/nd6.c 1.146
105usr.sbin/ndp/ndp.c 1.42 105usr.sbin/ndp/ndp.c 1.42
106 106
107 Instead of voodo casts use simple byte pointer arithmetic and memcpy to 107 Instead of voodo casts use simple byte pointer arithmetic and memcpy to
108 create the "packed" binary format we pass out to userland when querying 108 create the "packed" binary format we pass out to userland when querying
109 the router/prefix list. 109 the router/prefix list.
110 Simplify code to print the router/prefix list: use memcpy and local 110 Simplify code to print the router/prefix list: use memcpy and local
111 structs properly aligned on the stack to decode the binary format 111 structs properly aligned on the stack to decode the binary format
112 passed by the kernel - instead of (bogusly) assuming the format will 112 passed by the kernel - instead of (bogusly) assuming the format will
113 obey all local alignement requirements. 113 obey all local alignement requirements.
114 [martin, ticket #998] 114 [martin, ticket #998]
115 115
116sys/compat/common/compat_util.c 1.45 116sys/compat/common/compat_util.c 1.45
117 117
118 Free pathbuf in an error path. 118 Free pathbuf in an error path.
119 [martin, ticket #999] 119 [martin, ticket #999]
120 120
121sbin/fsck_ffs/inode.c 1.70 121sbin/fsck_ffs/inode.c 1.70
122 122
123 Fix cut-and-paste error in the non-ufs2 case which can cause 123 Fix cut-and-paste error in the non-ufs2 case which can cause
124 fsck_ffs to exit with an internal error. 124 fsck_ffs to exit with an internal error.
125 [bouyer, ticket #991] 125 [bouyer, ticket #991]
126 126
127sys/arch/xen/xen/evtchn.c 1.70 127sys/arch/xen/xen/evtchn.c 1.70
128 128
129 Remove the "evtchn_do_event: handler %p didn't lower ipl %d %d\n" 129 Remove the "evtchn_do_event: handler %p didn't lower ipl %d %d\n"
130 printf as analysis shows it actually isn't a bug in the handler, 130 printf as analysis shows it actually isn't a bug in the handler,
131 but related to spin mutexes. Fixes port-xen/46313. 131 but related to spin mutexes. Fixes port-xen/46313.
132 [bouyer, ticket #995] 132 [bouyer, ticket #995]
133 133
 134distrib/ews4800mips/Makefile 1.3
 135
 136 Add "cdroms" to the SUBDIR list for src/distrib/ews4800mips.
 137 Should prevent the iso image from being created in the source tree
 138 [apb, ticket #1009]
 139
 140etc/ntp.conf 1.16, 1.17, 1.18 via patch
 141external/bsd/ntp/dist/ntpd/ntp_request.c patch
 142
 143 Patch from ntp 4.2.7p404 to prevent an amplifier and DoS attack.
 144 Add several "restrict" lines to the default ntp.conf and
 145 improve comments
 146 [spz, ticket #1010]
 147