Re: Undocumented tape devices in pax(1)
On Mon, 8 May 2000, Warner Losh wrote: Leaving aside the 'r' question for the moment... Should that be sa or ast? sa is the scsi device for any tape device (formerly st or mt), while ast is for ide/atapi based tape drives. It should be ssa and asa, of course :-). The wt and wst devices referenced in our man pages are just plain bogus. I think we've killed all ft references in the tree... No, wst is still used by pc98, and wt is the Wangtek tape driver. wst and ast are weird names. Doesn't the "s" in them stand for "SCSI" and not "streaming", so wst is the so-called-Winchester (non-SCSI) SCSI tape driver, etc? For completeness, we should have had nrrrwsst (the non-rewinding rewinding raw so-called-Winchester streaming SCSI tape driver) ;-). Seriously, why aren't there "n" and "e" forms of ast? Bruce To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Tue, May 09, 2000 at 08:54:50PM +1000, Bruce Evans wrote: wst and ast are weird names. Doesn't the "s" in them stand for "SCSI" and not "streaming", so wst is the so-called-Winchester (non-SCSI) SCSI It does to me. But McKusick's mail I forwarded says "s" was for "streaming". -- -- David([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
The Pyramid series of machines used to have block tape devices, such that one was able to boot a repair kernel and ro root fs off the 1600bpi reel-to-reel deck. Not unaturally, one was discouraged from doing a recursive find on that fs. Stephen (who used to have thoughts of doing the same with his old QIC-150) -- The views expressed above are not those of PGS Tensor. "We've heard that a million monkeys at a million keyboards could produce the Complete Works of Shakespeare; now, thanks to the Internet, we know this is not true."Robert Wilensky, University of California -- The views expressed above are not those of PGS Tensor. "We've heard that a million monkeys at a million keyboards could produce the Complete Works of Shakespeare; now, thanks to the Internet, we know this is not true."Robert Wilensky, University of California To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Sun, May 07, 2000 at 02:56:17PM -0700, Mike Smith wrote: We had this argument the other day, and you clearly didn't understand. Yes I did. We agreed to not agree and to not argue it. :-) Which is why I've never brought it up with you again. -- -- David([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
In message [EMAIL PROTECTED] "David O'Brien" writes:
: On Sun, May 07, 2000 at 04:39:16PM +0200, Christian Weisgerber wrote:
: OpenBSD only changed "rmt" to "rst" ("rsa" for us)
:
: Just "sa" for us -- "sa" is now a raw device and "rFOO" use is
: depreciated.
Leaving aside the 'r' question for the moment...
Should that be sa or ast? sa is the scsi device for any tape device
(formerly st or mt), while ast is for ide/atapi based tape drives.
The wt and wst devices referenced in our man pages are just plain
bogus. I think we've killed all ft references in the tree...
Warner
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
In message [EMAIL PROTECTED] Matthew Jacob writes: : Oh, and in the updating of this, don't forget the FreeBSD usage of .ctl for : tape devices- as far as I know this is the only *BSD that has this. Which devices use .ctl? sa and ast don't seem to use them now (at the very least they aren't created by MAKEDEV by default). Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
In the last episode (May 08), Warner Losh said: In message [EMAIL PROTECTED] Matthew Jacob writes: : Oh, and in the updating of this, don't forget the FreeBSD usage of : .ctl for tape devices- as far as I know this is the only *BSD that : has this. Which devices use .ctl? sa and ast don't seem to use them now (at the very least they aren't created by MAKEDEV by default). *.ctl is handy for getting status on a device that another process has open; if I'm dumping to /dev/nrsa0, I can run "mt -f /dev/rsa0.ctl status" on another tty and see what file/block position the tape is at. Dunno if it has any other use :) -- Dan Nelson [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Mon, May 08, 2000 at 15:42:01 -0600, Warner Losh wrote: In message [EMAIL PROTECTED] Matthew Jacob writes: : Oh, and in the updating of this, don't forget the FreeBSD usage of .ctl for : tape devices- as far as I know this is the only *BSD that has this. Which devices use .ctl? sa and ast don't seem to use them now (at the very least they aren't created by MAKEDEV by default). The sa driver does use the control nodes, whether or not they're actually created by MAKEDEV. (Look in saopen().) It's useful to be able to get status on your tape drive while a backup is going on via the control node, e.g.: mt -f /dev/rsa0.ctl status Ken -- Kenneth Merry [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
In message [EMAIL PROTECTED] "Kenneth D. Merry" writes: : The sa driver does use the control nodes, whether or not they're actually : created by MAKEDEV. (Look in saopen().) It's useful to be able to get : status on your tape drive while a backup is going on via the control node, : e.g.: : : mt -f /dev/rsa0.ctl status OK. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Mon, 8 May 2000, Warner Losh wrote:
In message [EMAIL PROTECTED] Matthew Jacob
writes:
: Oh, and in the updating of this, don't forget the FreeBSD usage of .ctl for
: tape devices- as far as I know this is the only *BSD that has this.
Which devices use .ctl? sa and ast don't seem to use them now (at the
very least they aren't created by MAKEDEV by default).
Should be:
sa*)
umask $tape_umask
unit=`expr $i : '..\(.*\)'`
chr=14
case $unit in
[0-9]|[1-9][0-9]|[1-9][0-9][0-9]|[1-9][0-9][0-9][0-9])
mknod rsa${unit}.ctl c $chr `saminor 1 ${unit} 0 0`
for m in 0 1 2 3
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Mon, 8 May 2000, Dan Nelson wrote: In the last episode (May 08), Warner Losh said: In message [EMAIL PROTECTED] Matthew Jacob writes: : Oh, and in the updating of this, don't forget the FreeBSD usage of : .ctl for tape devices- as far as I know this is the only *BSD that : has this. Which devices use .ctl? sa and ast don't seem to use them now (at the very least they aren't created by MAKEDEV by default). *.ctl is handy for getting status on a device that another process has open; if I'm dumping to /dev/nrsa0, I can run "mt -f /dev/rsa0.ctl status" on another tty and see what file/block position the tape is at. Dunno if it has any other use :) Eventually it will have/set more extended error information. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Undocumented tape devices in pax(1)
Can someone explain to me why pax(1) has (undocumented) switches which select some tape devices, but apparently randomly numbered ones: From tar.h: /* * default device names */ #define DEV_0 "/dev/rmt0" #define DEV_1 "/dev/rmt1" #define DEV_4 "/dev/rmt4" #define DEV_5 "/dev/rmt5" #define DEV_7 "/dev/rmt7" #define DEV_8 "/dev/rmt8" These are selectable through -0, -1, -4, etc. Nevermind the fact that they point to devices which have never existed in FreeBSD, but why on earth wouldn't you want -2, -3, or -6? Anyway, does anyone see the point in leaving these in (changing the devices to /dev/rsa# and documenting their existence), or should I rip them out? Kris In God we Trust -- all others must submit an X.509 certificate. -- Charles Forsythe [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
Kris Kennaway [EMAIL PROTECTED] wrote:
Can someone explain to me why pax(1) has (undocumented) switches which
select some tape devices, but apparently randomly numbered ones:
Note that these switches appear only in pax' tar compatibility
personality, which isn't used in FreeBSD. And the reason they're
there is because old versions of tar had them.
I'm looking at 4.3BSD's usr/src/bin/tar.c right now, and it supports
-[014578] to select the respective "/dev/rmt#" device.
These are selectable through -0, -1, -4, etc. Nevermind the fact that they
point to devices which have never existed in FreeBSD, but why on earth
wouldn't you want -2, -3, or -6?
Historical reasons.
Back in 4.3BSD, bits 0 and 1 of the minor number seem to have
specified the device, bit 2 marked non-rewinding, bit 3 6250bpi
(high density?). That explains rmt[0145] well enough, although
rmt[78] remain unclear.
Anyway, does anyone see the point in leaving these in (changing the
devices to /dev/rsa# and documenting their existence), or should I rip
them out?
OpenBSD only changed "rmt" to "rst" ("rsa" for us), which isn't
particularly useful. Solaris uses 0..7 to select an entry from
/etc/default/tar, which specifies device name, block size, and tape
size.
I guess mapping -[01] to rsa[01], and -[45] to nrsa[01] still makes
about the most sense.
Unless you intend to revive pax' tar personality under FreeBSD
(which would suggest merging in OpenBSD's changes), I'd say just
leave it.
--
Christian "naddy" Weisgerber [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Sun, May 07, 2000 at 04:39:16PM +0200, Christian Weisgerber wrote:
OpenBSD only changed "rmt" to "rst" ("rsa" for us)
Just "sa" for us -- "sa" is now a raw device and "rFOO" use is
depreciated.
--
-- David([EMAIL PROTECTED])
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Sun, May 07, 2000 at 04:39:16PM +0200, Christian Weisgerber wrote:
OpenBSD only changed "rmt" to "rst" ("rsa" for us)
Just "sa" for us -- "sa" is now a raw device and "rFOO" use is
depreciated.
We had this argument the other day, and you clearly didn't understand.
We have three devices for each tape drive: rsaX, ersaX and nrsaX.
The 'r' prefix for tape devices is entirely unrelated to the 'r' prefix
for disk devices. Please read the sa(4) manpage prior to its' corruption
by a well-meaning but entirely ignorant committer. (Fixed, btw.)
--
\\ Give a man a fish, and you feed him for a day. \\ Mike Smith
\\ Tell him he should learn how to fish himself, \\ [EMAIL PROTECTED]
\\ and he'll hate you for a lifetime. \\ [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
Mike Smith [EMAIL PROTECTED] wrote: The 'r' prefix for tape devices is entirely unrelated to the 'r' prefix for disk devices. I'd like to see some backup for this assertion. Historically, BSD (up to 4.4) used to have mt block device, rewinding nmt block device, non-rewinding rmt character device, rewinding nrmt character device, non-rewinding which leaves little room to doubt. -- Christian "naddy" Weisgerber [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
Mike Smith [EMAIL PROTECTED] wrote: The 'r' prefix for tape devices is entirely unrelated to the 'r' prefix for disk devices. I'd like to see some backup for this assertion. Historically, BSD (up to 4.4) used to have mt block device, rewinding nmt block device, non-rewinding rmt character device, rewinding nrmt character device, non-rewinding which leaves little room to doubt. Interesting. I've never encountered a tape device to which the buffer cache was applicable, except for the mythical SunOS swap-to-tape story, and the 'r' has always been "rewind" for as long as I can remember. We haven't made non-'r' devices since MAKEDEV rev 1.5. -- \\ Give a man a fish, and you feed him for a day. \\ Mike Smith \\ Tell him he should learn how to fish himself, \\ [EMAIL PROTECTED] \\ and he'll hate you for a lifetime. \\ [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
On Sunday, 7 May 2000 at 16:48:36 -0700, Mike Smith wrote: Mike Smith [EMAIL PROTECTED] wrote: The 'r' prefix for tape devices is entirely unrelated to the 'r' prefix for disk devices. I'd like to see some backup for this assertion. Historically, BSD (up to 4.4) used to have mt block device, rewinding nmt block device, non-rewinding rmt character device, rewinding nrmt character device, non-rewinding which leaves little room to doubt. Interesting. I've never encountered a tape device to which the buffer cache was applicable, except for the mythical SunOS swap-to-tape story, and the 'r' has always been "rewind" for as long as I can remember. I suspect that's an assumption on your part. I think we've come up with enough man pages to support naddy's statement. We haven't made non-'r' devices since MAKEDEV rev 1.5. For good reasons :-) Greg -- Finger [EMAIL PROTECTED] for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Undocumented tape devices in pax(1)
Greg Lehey [EMAIL PROTECTED] wrote: I suspect that's an assumption on your part. I think we've come up with enough man pages to support naddy's statement. Which, btw, was drawn from inspection of MAKEDEV in the various 4.xBSD releases in the CSRG archives (Kirk's CD set). Personally, I take that as an authoritative source. -- Christian "naddy" Weisgerber [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
