+ The bufov vulnerability was fixed in the vendor sources.
+ The vendor sources are now POSIX [me harder] compliant.
+ The db vs. dbm files issue was fixed in the vendor sources.
__swrite() and __sseek() to higher level. According to funopen(3) they all
are just wrappers to something like standard read(2), write(2) and
lseek(2), i.e. must not touch stdio internals because they are replaceable
with any other functions knows nothing about stdio internals. See example
of funopen(3) usage in sendmail sources f.e.
NOTE: this is original stdio bug, not result of my range checkin added.
I guess I pooched the permissions on the scripts before committing them
since they're not executable and now it's too late to change (I think -
I suppose you could chmod the ,v files and it might inheirit but I don't
feel like asking the repomeisters to try that).
Noticed by: Dirk Froemberg <dirk@freebsd.org>
I don't recall why the rev 1.2 hack is needed. It looks like Amd was
already using our local headers. I no longer have a FreeBSD 2.x box to
test this on, and this will never be MFC'ed to RELENG_2_2. So lets just
decrease the maintenance effort.
+ The rev 1.4 addition was taken from a snapshot previous to 6.0.7, so
it is included in 6.0.7.
+ The vendor sources are now POSIX [me harder] compliant.
-O, which limits the impact of the write-only restriction to guest
users.
*) The existing manual page's SYNOPSIS and option listing in the
DESCRIPTION are already horribly disordered. No attempt has been
made to fix this.
*) The existing source's getopt() optstring and option handling switch
are already horribly disordered. No attempt has been made to fix
this.
Discussed with: nik, -audit
fseek -> fseeko
ftell -> ftello
NOTE: fseek/ftell not works for >long offsets per POSIX:
[EOVERFLOW] For fseek( ), the resulting file offset would be a value which
cannot be represented correctly in an object of type long.
[EOVERFLOW] For ftell ( ), the current file offset cannot be represented
correctly in an object of type long.
long -> time_t
%ld -> %qd
fseek -> fseeko
NOTE: that fseek not works for >long offsets per POSIX:
[EOVERFLOW] For fseek( ), the resulting file offset would be a value which
cannot be represented correctly in an object of type long.
A nsp chip does suspend I/O write by 512bytes burst write,
though the chip only has 48 bytes FIFO. The chip assert I/O WAIT
signal to PC-Card bus after the CPU writes more than 48 bytes to
the chip if the SCSI device does not respond immediately in supsend
I/O burst write. If the device does not respond for a while it might
cause PC-Card bus timeout.
The previous work around was to wait the request from SCSI device.
But there are some devices which request bytes for synchronous transfer
immediately. So current work aound is to fill 32bytes FIFO, wait for
FIFO empty and burst write 512-32 bytes for every 512 bytes block.
Submitted-by: Honda-san (the author of the driver)
Obtained-from: NetBSD/pc98
