Annotation of src/PATCHLEVEL, Revision 1.1.2.16
1.1.2.1 mycroft 1: This file describes official patches for the NetBSD 1.1 release.
2:
1.1.2.2 mycroft 3:
4: Mondo patch 1:
1.1.2.1 mycroft 5:
6: [kernel/i386]: Fix a compilation error when no DMA devices are
7: configured.
8:
1.1.2.3 mycroft 9: [kernel/i386]: Fix a problem where processes trying to use the FPU for
10: the first time would sporadically core dump. Known to affect httpd
11: and gopherd, in particular.
1.1.2.1 mycroft 12:
13: [kernel/i386]: Fix an oversight which caused the Ultrastor 24f probe
14: to always fail.
15:
16: [kernel/generic]: Fix a sgtty emulation problem which caused LITOUT
17: and PASS8 modes to sometimes be set erroneously (sometimes referred to
18: as `stair-stepping xterms').
19:
20: [kernel/generic]: Fix a problem in FFS that could cause panics in
21: ffs_clusteracct() after running fsck(8).
1.1.2.4 mycroft 22:
23:
24: Mondo patch 2:
25:
1.1.2.5 mycroft 26: [kernel/i386]: Redo the way the GDT is allocated, to avoid a potential
27: reentrance problem.
28:
1.1.2.4 mycroft 29: [kernel/generic]: Fix a reversed test in the tty driver that caused
30: ECHONL to not work correctly.
31:
32: [kernel/generic]: Fix a problem in many system calls which take a file
33: descriptor as an argument, which could cause the system to crash.
34:
35: [kernel/generic]: Speed up searching for a TCP control block by
36: hashing certain values.
37:
38: [kernel/generic]: Fix a bad interaction with the Linux NFS server by
39: not specifying an explicit gid (or uid) in create and related calls.
40:
41: [kernel/generic]: Fix problems in TCP that could, in strange
42: circumstances, cause sockets to be left in CLOSE_WAIT or FIN_WAIT_2
43: states.
44:
45: [kernel/generic]: Fix a problem in bind(2) that could allow a user to
46: steal a port from another process.
47:
48: [user/generic]: Fix oversights in quota-related utilities that would
49: cause them to not work on file systems with type `ffs' in fstab.
1.1.2.6 gwr 50:
51: [kernel/sun3]: Make netboot compatible with old Sun3/50 PROMs
52:
53: [kernel/sun3]: Try to deal with Sun3/50 SCSI DMA hardware bug
1.1.2.7 thorpej 54:
55: [kernel/hp300]: Fix incorrect value in HP-IB command stream.
1.1.2.8 thorpej 56:
57: [kernel/sparc]: Fix panic if more than one framebuffer is found
58: on a sun4 system.
59:
60: [kernel/sparc]: Fix oversight that caused rconsole to not attach to
61: cgsix-based consoles.
1.1.2.9 thorpej 62:
63: [kernel/hp300]: Fix integer overflow in page table map allocation that
64: could cause systems to hang during fork().
1.1.2.10 hpeyerl 65:
66: [user/generic]: Fix rarpd so that it only opens one BPF per interface
67: and not per address. Fixes problem where interface aliases caused
68: rarpd to open extra BPF's.
1.1.2.11 scottr 69:
70: [kernel/i386]: Fix register constraints so gcc-2.7.2 can be used to
71: build kernels.
1.1.2.12 is 72:
73: [amiga]: Fix documentation about CPU support: we don't support 68060, only
74: 68020, 68030 and 68040.
75:
1.1.2.13 is 76: [kernel/amiga]: Fix a problem with the A2410 driver on fast CPUs and another
77: one which caused it to reject the last monitor definition.
78:
1.1.2.15 is 79: [kernel/amiga]: The "sfas" driver uses by the Fastlane and Blizzard SCSI
80: boards would attempt to write a byte register with a word
1.1.2.14 is 81: value, leading to wrong transfer counts in certain situations. Fix this.
1.1.2.15 is 82:
83: [kernel/amiga]: Fix the Cybervision driver, which used to fail on boards with
84: new revisions of the S3 chipset.
1.1.2.14 is 85:
1.1.2.16! is 86: [kernel/generic]: Fix the hardware independent ARCnet driver, which
! 87: could cause the kernel to panic if the interface was ifconfig'd down
! 88: right before the IP layer tried to output through it.
! 89:
! 90:
CVSweb <webmaster@jp.NetBSD.org>