Re: ELF rtld and environment variables...
Ollivier Robert wrote: According to Julian Stacey: 4.1-release produces no /sbin/mount_cfs, man mount give no hint, If you have patches to test, I volunteer to test on 4.1 or 3.4 :-) It is a port. I'd love to import it into CURRENT though. Some friends running vile Micro$oft asked me if BSD offers an encrypting file system, it would be just too horrible to say "No", [though wether src/ or ports/ is best, I'm not now informed to comment] How do I get my hands on your sources ? :-) I'm running 4.0 on my laptop, was going to 4.1, but will go stable or current instead if necessary. Julian - Julian Stacey http://bim.bsn.com/~jhs/ Munich Unix Consultant. Free BSD Unix with 3600 packages sources. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
Julian Stacey wrote:
Ollivier Robert wrote:
According to Julian Stacey:
4.1-release produces no /sbin/mount_cfs, man mount give no hint,
If you have patches to test, I volunteer to test on 4.1 or 3.4 :-)
It is a port. I'd love to import it into CURRENT though.
Some friends running vile Micro$oft asked me if BSD offers an encrypting file
system, it would be just too horrible to say "No",
[though wether src/ or ports/ is best, I'm not now informed to comment]
How do I get my hands on your sources ? :-) I'm running 4.0 on my laptop,
was going to 4.1, but will go stable or current instead if necessary.
My relatively recent 4.1 laptop has it in ports/security/cfs. The package
description reads:
This is CFS, Matt Blaze's Cryptographic File System. It provides
transparent encryption and decryption of selected directory trees.
It is implemented as a user-level NFS server and thus does not
require any kernel modifications.
For an overview of how to use it, read "${PREFIX}/share/doc/cfs/notes.ms"
and the manual pages. There is a paper describing CFS at:
ftp://research.att.com/dist/mab/cfs.ps
Under FreeBSD, the mount command for the CFS tree must include
"-o port=3049,nfsv2".
John Polstra [EMAIL PROTECTED]
--
"Where am I, and what am I doing in this handbasket?"
Wes Peters Softweyr LLC
[EMAIL PROTECTED] http://softweyr.com/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
Ollivier Robert wrote: According to Julian Stacey: just as today I'd use an encrypting file system on my new laptop, but such file system don't exist on FreeBSD unfortunately. Ahem. Why did I sent an update for security/cfs to green a few months ago? :-) 4.1-release produces no /sbin/mount_cfs, man mount give no hint, If you have patches to test, I volunteer to test on 4.1 or 3.4 :-) Julian - Julian Stacey http://bim.bsn.com/~jhs/ Munich Unix Consultant. Free BSD Unix with 3600 packages sources. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
According to Julian Stacey: 4.1-release produces no /sbin/mount_cfs, man mount give no hint, If you have patches to test, I volunteer to test on 4.1 or 3.4 :-) It is a port. I'd love to import it into CURRENT though. -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- [EMAIL PROTECTED] FreeBSD keltia.freenix.fr 5.0-CURRENT #80: Sun Jun 4 22:44:19 CEST 2000 To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
On Wed, 26 Jul 2000, Julian Stacey wrote: That laptop has now gone to 4.0, aout to elf, a 1.5G disc, so no incentive to do it all again to see how much FreeBSD-4 gzipped aout binary tree might save/waste on a whole tree. BTW I was `strip'ing gzexe(1) is your friend :-) 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-hackers" in the body of the message
Re: ELF rtld and environment variables...
"Andrew Reilly" wrote: Well, even if there are/were folk who want tiny disk footprints, and crunching everything isn't going to do the whole job, wouldn't a compressed filesystem be a better way to approach this? At least that way you'd still be able to page from the executable(s), and all of the on-disk data would bennefit too. I would have used a compressing file system if it had existed, just as today I'd use an encrypting file system on my new laptop, but such file system don't exist on FreeBSD unfortunately. Nice ideas to add to a web page of project ideas for students other contributors. When I was a student, friends were keen to find project ideas more inspiring than those on lists drawn up by the lecturers. Julian - Julian Stacey http://bim.bsn.com/~jhs/ Munich Unix Consultant. Free BSD Unix with 3200 packages sources. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
According to Julian Stacey: just as today I'd use an encrypting file system on my new laptop, but such file system don't exist on FreeBSD unfortunately. Ahem. Why did I sent an update for security/cfs to green a few months ago? :-) -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- [EMAIL PROTECTED] FreeBSD keltia.freenix.fr 5.0-CURRENT #80: Sun Jun 4 22:44:19 CEST 2000 To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
From: Mike Smith [EMAIL PROTECTED] Mike Smith wrote: gzipped binaries are actually a terrible idea; they actually *waste* space in most cases. Suprising, They saved space for a 200M disc in a 486 laptop with 3.[2,3,or4], it was so tight for space I gzipped everything, (entire output of src/ except kernel), to save space to then install X/ or something else big. That laptop has now gone to 4.0, aout to elf, a 1.5G disc, so no incentive to do it all again to see how much FreeBSD-4 gzipped aout binary tree might save/waste on a whole tree. BTW I was `strip'ing everything, did not use more than standard optimiser, didnt compile with -g. Script I used is on http://bim.bsn.com/~jhs/bin/.csh/gzip_bins Julian - Julian Stacey http://bim.bsn.com/~jhs/ Kostenlos: FreeBSD 3200 packages, sources, Netscape, WordPerfect StarWriter. Anti Software Patent Petition to Euro Parliament: http://petition.eurolinux.org To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
From: Mike Smith [EMAIL PROTECTED] Mike Smith wrote: gzipped binaries are actually a terrible idea; they actually *waste* space in most cases. Suprising, They saved space for a 200M disc in a 486 laptop with 3.[2,3,or4], No, that's the one case where they help. But people aren't trying to squeeze whole systems into small disks anymore; they're trying to run cut-down systems in tiny spaces (where the fact that you have to unpack the entire binary into memory hurts), or disk space is so cheap that the speed/swap hit is the only impacting factor. Typically, the loss of the ability to demand-page from a gzipped executable is a worse detracting factor than the space saving makes up for. -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
Mike Smith wrote: gzipped binaries are actually a terrible idea; they actually *waste* space in most cases. Suprising, They saved space for a 200M disc in a 486 laptop with 3.[2,3,or4], No, that's the one case where they help. But people aren't trying to squeeze whole systems into small disks anymore; Really? News to me... they're trying to run cut-down systems in tiny spaces (where the fact that you have to unpack the entire binary into memory hurts), or disk space is so cheap that the speed/swap hit is the only impacting factor. Methinks you generalize *way* too much, w/out knowing all the facts. Nate To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
In message [EMAIL PROTECTED] Mike Smith writes: : Typically, the loss of the ability to demand-page from a gzipped : executable is a worse detracting factor than the space saving makes up : for. This is one reason that Timing Solutions runs all of its small systems out of uncompressed flash or Disk On Chip. We thought about saving a little money and going to 8MB or 16MB parts rather than 64M parts and decompressing into memory, but some simple exeriments that I did showed that this savings would be offset by needing more RAM to hold the entire decompressed image. Much of the stuff we have on our CF parts is needed only for boot or configuration, so it turns out that only a small part of the binaries need to be in RAM at any given time. The price difference between either the 8MB or 16M and 64M CF part is something like $80. The price for an additional 32MB of ram is about $70-$80, give or take in the 72pin SIMMs that we have on our boards. Given our volumes, it didn't make sense to spend the 50 or so hours needed to make a compressed solution bulletproof, plus the unknown amount of time that a compressed solution causes for updates and patches. Even if we did 100 units, that's only $1000, which, counting overhead, is like 5-10 hours of somebody's time, which is far below the cost it would take to engineer the solution, plus build process additions, upgrade hassles, etc. The .gz files would solve only the upgrade hassle issue, while leaving the other issues in place. Each company has its own cost/benefit analysis for these things. So far none of them have seen enough of a benefit to .gz executables to implement them for elf. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
On Wed, Jul 26, 2000 at 02:00:46PM -0600, Nate Williams wrote: No, that's the one case where they help. But people aren't trying to squeeze whole systems into small disks anymore; Really? News to me... Well, even if there are/were folk who want tiny disk footprints, and crunching everything isn't going to do the whole job, wouldn't a compressed filesystem be a better way to approach this? At least that way you'd still be able to page from the executable(s), and all of the on-disk data would bennefit too. (I've read serious suggestions in comp.arch that it could be benneficial to compress DRAM, with hardware decompression/ compression on the way in and out of cache...) -- Andrew To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
On Mon, 24 Jul 2000, John Polstra wrote: Glad you liked the idea! :-) Well imagine if Joe user gets a Linux binary or a.out binary to run. Bam, it doesn't run, and one'd have to check each file, and unset the variables. Or forgo any user-feedback. :( Well, there is a different reason for each of the dynamic linkers. FreeBSD ELF: It's required by the ELF specification. FreeBSD a.out: Backward compatibility. Linux ELF: Because it's part of Linux and that's just what it does. I can understand the logic for using said variables for FreeBSD ELF stuff, but for the rest of them, I figure we're not actually the native environment. Hmm. Anywho the topic of caching shlib symbols came up in discussion as a possible way speed up loading of programs. Makes me wonder if it would be worth it.. - alex To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
In article [EMAIL PROTECTED], Alex Zepeda [EMAIL PROTECTED] wrote: On Mon, 24 Jul 2000, John Polstra wrote: FreeBSD ELF: It's required by the ELF specification. FreeBSD a.out: Backward compatibility. Linux ELF: Because it's part of Linux and that's just what it does. I can understand the logic for using said variables for FreeBSD ELF stuff, but for the rest of them, I figure we're not actually the native environment. Hmm. That's why we can't change the current behavior. The Linux RTLD is a 3rd party program. We didn't write it, and we don't maintain it. And I'm not about to break backward compatibility in the a.out RTLD. Anywho the topic of caching shlib symbols came up in discussion as a possible way speed up loading of programs. Makes me wonder if it would be worth it.. The only way to find out would be to take some measurements. But where were you thinking of caching these symbols? You can't do it in the RTLD's address space because the RTLD isn't persistent. You get a new instance of it every time you run a program. And if you want to cache the symbols in a file, well, we already have that -- it's called the shared library's symbol table. I really don't see any potential for gains from this. Besides, have you even established that dynamically linked programs load too slowly? I've certainly never heard any complaints along those lines. Furthermore I would bet that the bulk of the dynamic linking time comes from opening the shared libraries and mmapping them, and there's nothing the RTLD can do about that. John, who spent weeks instrumenting and measuring the a.out RTLD years ago. -- John Polstra [EMAIL PROTECTED] John D. Polstra Co., Inc.Seattle, Washington USA "Disappointment is a good sign of basic intelligence." -- Chögyam Trungpa To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
Besides, have you even established that dynamically linked programs load too slowly? I've certainly never heard any complaints along those lines. Furthermore I would bet that the bulk of the dynamic linking time comes from opening the shared libraries and mmapping them, and there's nothing the RTLD can do about that. John, who spent weeks instrumenting and measuring the a.out RTLD years ago. And optimized it greatly after these weeks, in one of his first projects in FreeBSD. :) :) :) Nate To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
From: John Polstra [EMAIL PROTECTED] No, there isn't. I don't plan to do anything more with the a.out dynamic linker, as I consider it obsolete at this point. I'd BTW (last I looked) support of gzipped execs was only available for aout, not for elf, ... one more residual use for aout, (apart from the discussed default netscape) Julian - Julian Stacey http://bim.bsn.com/~jhs/ Anti Software Patent Petition to Euro Parliament: http://petition.eurolinux.org To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
From: John Polstra [EMAIL PROTECTED] No, there isn't. I don't plan to do anything more with the a.out dynamic linker, as I consider it obsolete at this point. I'd BTW (last I looked) support of gzipped execs was only available for aout, not for elf, ... one more residual use for aout, (apart from the discussed default netscape) gzipped binaries are actually a terrible idea; they actually *waste* space in most cases. -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
In article [EMAIL PROTECTED], Alex Zepeda [EMAIL PROTECTED] wrote: [LD_PRELOAD and LD_LIBRARY_PATH] This works great if you're using your average ELF binary. However, if you're like me and still somewhat attached to the idea of using Navigator which is an a.out binary (perhaps the only one I still have left), you're SOL as the variables tweak the a.out RTLD too. This perhaps doesn't sound too bad, until one realizes that the said shlib objets are ELF, and the a.out rtld doesn't grok any of them, meaning that one has to fire up an xterm, unset the said variables to start netscape. My question is this: is there any way to make the a.out rtld ignore these variables? No, there isn't. I don't plan to do anything more with the a.out dynamic linker, as I consider it obsolete at this point. I'd suggest making a script "run_aout" that looks something like this (untested): #! /bin/sh unset LD_PRELOAD unset LD_LIBRARY_PATH exec $0 "$@" BTW, it's generally not a good idea to set either of those environment variables globally. You should only set them in scripts which run specific executables that need them to be set. Besides the a.out problem, they affect programs run under Linux emulation too. John -- John Polstra [EMAIL PROTECTED] John D. Polstra Co., Inc.Seattle, Washington USA "Disappointment is a good sign of basic intelligence." -- Chögyam Trungpa To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
On Mon, Jul 24, 2000 at 03:25:07AM -0700, Alex Zepeda wrote: you're like me and still somewhat attached to the idea of using Navigator which is an a.out binary (perhaps the only one I still have left), you're Use the BSDI Netscape ports, which are ELF and don't require any emulation. They are also fast and seldom crash. IMNSHO, the FreeBSD Netscape ports should be axed if Netscape won't ever make ELF binaries. -- Will Andrews [EMAIL PROTECTED] [EMAIL PROTECTED] GCS/E/S @d- s+:+ a--- C++ UB$ P+ L- E--- W+ N-- !o ?K w--- O- M+ V- PS+ PE++ Y+ PGP t++ 5 X+ R+ tv+ b++ DI+++ D+ G++ e h! r- y? To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
On Mon, 24 Jul 2000, John Polstra wrote: No, there isn't. I don't plan to do anything more with the a.out dynamic linker, as I consider it obsolete at this point. I'd suggest making a script "run_aout" that looks something like this (untested): Uck. BTW, it's generally not a good idea to set either of those environment variables globally. You should only set them in scripts which run specific executables that need them to be set. Besides the a.out problem, they affect programs run under Linux emulation too. Yes, but this is intended to be used for all programs (not an awful idea IMO, questionable implementation, but what other alternatives are there?). It's designed to give the user feedback that an application has been started so one doesn't double click on an icon multiple times. I'm curious, why do a.out/FreeBSD-elf/Linux-elf programs all respond to the same variables? Sure it's perhaps a consistant interface, but wouldn't somthing like LINUX_LD_LIBRARY_PATH and/or AOUT_LD_LIBRARY_PATH make more sense? - alex To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: ELF rtld and environment variables...
In article [EMAIL PROTECTED], Alex Zepeda [EMAIL PROTECTED] wrote: Uck. Glad you liked the idea! :-) I'm curious, why do a.out/FreeBSD-elf/Linux-elf programs all respond to the same variables? Sure it's perhaps a consistant interface, but wouldn't somthing like LINUX_LD_LIBRARY_PATH and/or AOUT_LD_LIBRARY_PATH make more sense? Well, there is a different reason for each of the dynamic linkers. FreeBSD ELF: It's required by the ELF specification. FreeBSD a.out: Backward compatibility. Linux ELF: Because it's part of Linux and that's just what it does. John -- John Polstra [EMAIL PROTECTED] John D. Polstra Co., Inc.Seattle, Washington USA "Disappointment is a good sign of basic intelligence." -- Chögyam Trungpa To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
