Re: 80386 out of GENERIC
On Wed, Dec 18, 2002 at 03:33:42PM +1100, Bruce Evans wrote:
On Tue, 17 Dec 2002, Ruslan Ermilov wrote:
On Mon, Dec 16, 2002 at 09:05:40AM +1030, Greg 'groggy' Lehey wrote:
I suppose it would be a good idea to include an alternatvie i386
kernel on the CD-ROM. There may be a space issue, of course. How
many people participating in this thread have an i386 with at least 12
MB of memory and intended to try 5.0 on it? How many of those don't
have a machine to bootstrap off?
Having only alternative i386 kernel is not enough while userland
stuff is still compiled for i486.
Er, userland stuff is still compiled for original i386's, modulo bugs.
E.g., in the i386 endian.h:
% #if defined(_KERNEL) (defined(I486_CPU) || defined(I586_CPU) ||
defined(I686_CPU)) !defined(I386_CPU)
^^^
%
% #define __byte_swap_int(x) \
% __extension__ ({ register __uint32_t __X = (x); \
%__asm (bswap %0 : +r (__X)); \
%__X; })
% #else
%
% #define __byte_swap_int(x) \
% __extension__ ({ register __uint32_t __X = (x); \
%__asm (xchgb %h0, %b0\n\trorl $16, %0\n\txchgb %h0, %b0 \
% : +q (__X)); \
%__X; })
% #endif
The _KERNEL part of the ifdef limits the use of the i486 bswap
instruction to the kernel, so userland is properly pessimized to support
all x86's. The other parts of the ifdef properly pessimize modules to
support all x86's (options don't apply to modules so none of the XXX_CPU's
is defined).
So kernels get the full epsilon of optimizations from turning off i386
support, while userland doesn't get any (not counting ones from optimizing
for non-i386 without breaking i386).
Right. I forgot for the moment that -mcpu isn't the same as -march,
and the latter is only set if CPUTYPE is also set.
Cheers,
--
Ruslan Ermilov Sysadmin and DBA,
[EMAIL PROTECTED] Sunbay Software AG,
[EMAIL PROTECTED] FreeBSD committer,
+380.652.512.251Simferopol, Ukraine
http://www.FreeBSD.org The Power To Serve
http://www.oracle.com Enabling The Information Age
msg48985/pgp0.pgp
Description: PGP signature
Re: 80386 out of GENERIC
Leif Neland wrote: But still, would it be impossible to have both a GENERIC and a GENERIC386 kernel in the distribution? Impossible, no... but would anyone use it? Seems to me that it would just take up space. And it's one more thing for the build gurus to keep a configuration for (though maybe they don't mind/care?). As long as someone can build a 386 kernel by throwing a few flags... To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On 18-Dec-2002 Terry Lambert wrote:
John Baldwin wrote:
This has nothing to do with /dev/random. Please stop with the constant
FUDing Terry.
| Revision 1.296 / (download) - annotate - [select for diffs], Sun Jan 14
| 10:11:10 2001 UTC (23 months ago) by jhb
| Branch: MAIN
| Changes since 1.295: +2 -2 lines
| Diff to previous 1.295 (colored)
|
| Remove I386_CPU from GENERIC. Support for the 386 seriously pessimizes
| performance on other x86 processors. Custom kernels can still be built
| that will run on the 386.
The pessimization that was being discussed right before that happened
was harvesting entropy for /dev/random. I can provide mailing list
quotes about that bracketing those dates.
No, the pessimization is:
#if defined(I386_CPU) || defined(CPU_DISABLE_CMPXCHG)
static __inline int
atomic_cmpset_int(volatile u_int *dst, u_int exp, u_int src)
{
int res = exp;
__asm __volatile(
pushfl ;
cli ;
cmpl%0,%2 ;
jne 1f ;
movl%1,%2 ;
1:
sete%%al;
movzbl %%al,%0 ;
popfl ;
# atomic_cmpset_int
: +a (res)/* 0 (result) */
: r (src),/* 1 */
m (*(dst)) /* 2 */
: memory);
return (res);
}
#else /* defined(I386_CPU) */
static __inline int
atomic_cmpset_int(volatile u_int *dst, u_int exp, u_int src)
{
int res = exp;
__asm __volatile (
__XSTRING(MPLOCKED)
cmpxchgl %1,%2 ;
setz%%al ;
movzbl %%al,%0 ;
1:
# atomic_cmpset_int
: +a (res)/* 0 (result) */
: r (src),/* 1 */
m (*(dst)) /* 2 */
: memory);
return (res);
}
#endif /* defined(I386_CPU) */
Was there a particular pessimization other than /dev/random that you
were thinking of when you made the commit comment?
Yes, the one above. It came in as part of SMPng.
--
John Baldwin [EMAIL PROTECTED]http://www.FreeBSD.org/~jhb/
Power Users Use the Power to Serve! - http://www.FreeBSD.org/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
I can't believe this thread is still polluting the email system. 386's are old, slow, and virtually useless. I think that the time wasted on supporting junk hardware would be better spent on utilising the features and capabilities of new hardware. As someone mentioned, if you want to use crap hardware, install NetBSD. FreeBSD's goal isn't to be able to run on anything, it's to be able to run fast on specific things. With that in mind, put your 386's away, or find a different OS. I know it may be hard to part with old junk, but that's life. -Craig - Original Message - From: Leif Neland [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Wednesday, December 18, 2002 02:12 Subject: Re: 80386 out of GENERIC But still, would it be impossible to have both a GENERIC and a GENERIC386 kernel in the distribution? Or is the whole system compiled in non-386 mode? Even so, if just one site. www.386.freebsd.org were having a 386-enabled version available, wouldn't that make everybody happy? Leif To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Tue, 17 Dec 2002 19:47:24 -0800 Juli Mallett [EMAIL PROTECTED] wrote: * De: Terry Lambert [ Data: 2002-12-17 ] [ Subjecte: Re: 80386 out of GENERIC ] Apparently, one of the primary markets for FreeBSD is embedded devices. Are you implying that these people, who are undoubtedly adding and removing lots of things in the kernel, to make things fit, and to make things do their jobs, can't be bothered to use the appropriate CPU settings? I'm implying that they *can* be bothered. I'm also implying that they *are* bothered. Finally, I'm saying that the less bother, the better. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Tue, 17 Dec 2002 20:09:10 -0800 Juli Mallett [EMAIL PROTECTED] wrote: I don't think anyone wants that to happen (though I wouldn't put it past some people to want to do that). FWIW, all I've really been doing is defending the idea that some people want the code to run on a 386, however poorly. But for the most part, the people who would use that approach aren't bothered by not having 386 support there by default. BTW, for the person who pointed out that the userland is 486, and won't run on 386, anyway, without everything rebuilt: I'm pretty sure that's not true since the compiler upgrade, since the new compiler can't do CPU optimization correctly, without puking things out. 8-). -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
From: Craig Reyenga [EMAIL PROTECTED] I can't believe this thread is still polluting the email system. 386's are old, slow, and virtually useless. I think that the time wasted on supporting junk hardware would be better spent on utilising the features and capabilities of new hardware. As someone mentioned, if you want to use crap hardware, install NetBSD. FreeBSD's goal isn't to be able to run on anything, it's to be able to run fast on specific things. With that in mind, put your 386's away, or find a different OS. I know it may be hard to part with old junk, but that's life. I take great offense in all of what you posted. Your opinion is extremely short-sighted and one-sided. I don't have any problem with removing 386 support from GENERIC, and there need be no requirement to pre-build anything beyond what has already been decided, but I would hope that it will be possible to simply 'make' a 386 kernel--and all the rest--indefinitely. -- Paul A. Scott mailto:[EMAIL PROTECTED] http://skycoast.us/pscott/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Wed, 18 Dec 2002, John Baldwin wrote:
On 18-Dec-2002 Terry Lambert wrote:
The pessimization that was being discussed right before that happened
was harvesting entropy for /dev/random. I can provide mailing list
quotes about that bracketing those dates.
No, the pessimization is:
#if defined(I386_CPU) || defined(CPU_DISABLE_CMPXCHG)
static __inline int
atomic_cmpset_int(volatile u_int *dst, u_int exp, u_int src)
{
int res = exp;
__asm __volatile(
pushfl ;
cli ;
cmpl%0,%2 ;
jne 1f ;
movl%1,%2 ;
1:
sete%%al;
movzbl %%al,%0 ;
popfl ;
# atomic_cmpset_int
: +a (res)/* 0 (result) */
: r (src),/* 1 */
m (*(dst)) /* 2 */
: memory);
return (res);
}
#else /* defined(I386_CPU) */
static __inline int
atomic_cmpset_int(volatile u_int *dst, u_int exp, u_int src)
{
int res = exp;
__asm __volatile (
__XSTRING(MPLOCKED)
cmpxchgl %1,%2 ;
setz%%al ;
movzbl %%al,%0 ;
1:
# atomic_cmpset_int
: +a (res)/* 0 (result) */
: r (src),/* 1 */
m (*(dst)) /* 2 */
: memory);
return (res);
}
#endif /* defined(I386_CPU) */
Was there a particular pessimization other than /dev/random that you
were thinking of when you made the commit comment?
Yes, the one above. It came in as part of SMPng.
I think this is an insignificant pessimization (except for vmware). The
main issue is that the i386 version doesn't actually work for SMP. It
would need to be more comlicated and pessimal to work.
BTW, I have micro-optimized the above (to usually avoid the movzbl and
to allocate registers better, especially the hard register %al). One
reason that I haven't committed the changes is that I haven't got around
to testing them for the CPU_I386 case.
Bruce
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: 5.0 performance (was: 80386 out of GENERIC)
When I put KDE 3.0.5 on my 5.0-RC1 box it took just about 20 hours (600mhz with 384MB PC100... backup box ;)) but it runs faster (as compared to KDE on a 4.7-RELEASE). Im willing to give up a bit more comiple time to get better performance out of my apps. =) -Nick Harm Hale - Original Message - From: Cliff L. Biffle [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, December 17, 2002 1:29 AM Subject: Re: 5.0 performance (was: 80386 out of GENERIC) On Tuesday 17 December 2002 12:19 am, Garance A Drosihn wrote: At 5:58 AM +0100 12/17/02, Cliff Sarginson wrote: Also didn't someone mention that GCC has got slower anyway ? gcc is slower at compiling things. This is very noticeable when you're doing a buildworld. The code which gcc 3.2.1 produces does not seem any slower than the code produced by gcc 2.95.4 (the version in freebsd-stable). Actually, in my benchmarks here, the same code tends to yield much faster executables under gcc3, particularly in C++. But these are limited benchmarks (primarily of KDE and my own applications). I'm willing to trade some time on the compile (which, with any luck, happens only once) in exchange for speed in the application (which I may use every day)! :-) -Cliff L. Biffle To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 5.0 performance (was: 80386 out of GENERIC)
On 2002-12-16 23:24, Gary Stanley [EMAIL PROTECTED] wrote: At 03:45 AM 12/17/2002 +0200, Giorgos Keramidas wrote: I still have the Pentium 133 with 64 MB or memory that I used to run 5.0-CURRENT until a few weeks ago. I haven't got any real numbers, but the general `feel' of the system was pretty good. [...] Read the top of /usr/src/UPDATING Explains most of the slow problems. You got me backwards there. Removing all sort of debugging from current actually results in a fairly stable and fast system. It's the build of it all that is slow on an old machine (for obvious reasons, since FreeBSD is now a big system with a lot of tools). To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
John Baldwin wrote: This has nothing to do with /dev/random. Please stop with the constant FUDing Terry. | Revision 1.296 / (download) - annotate - [select for diffs], Sun Jan 14 | 10:11:10 2001 UTC (23 months ago) by jhb | Branch: MAIN | Changes since 1.295: +2 -2 lines | Diff to previous 1.295 (colored) | | Remove I386_CPU from GENERIC. Support for the 386 seriously pessimizes | performance on other x86 processors. Custom kernels can still be built | that will run on the 386. The pessimization that was being discussed right before that happened was harvesting entropy for /dev/random. I can provide mailing list quotes about that bracketing those dates. Was there a particular pessimization other than /dev/random that you were thinking of when you made the commit comment? The major functional changes immediately preceeding the disabling were 1.283, 1.279, and 1.275. 1.275 made /dev/random mandatory, and 1.283 disabled the blocking model /dev/random, to address hanges even on non-i386 architectures. 1.279 was Peter's cleanup, and doesn't seem to impact performance, even though it was moderately major. As I already pointed out: with the /dev/random algorithm now much more efficient than when it was first committed, maybe the impetus for axing i386 is no longer there. In any case, it can't hurt to periodically examine whether the reasons for the change are still valid or not, no matter what the pessimization was that was being referred to. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Chris Doherty wrote: p.s. I somehow suspect that embedded systems vendors aren't installing from the CDROM. why is this an issue? 1) supporting every computer made since 1964 is NetBSD's job, not FreeBSD's. 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. 3) if you feel compelled to run old hardware, why not shell out $30 for a 486 system? for $50 you can get a Pentium 166. :-) I'm really keen to see FreeBSD move *forward*. Apparently, one of the primary markets for FreeBSD is embedded devices. Macrocell libraries, from which CPU cores are assembled with purpose functional macrocells for embedded devices, such as the Apple AirPort (which is a 386 device), and similar devices, often offer 386's. Several vendors libraries offer 486's, but not all of them. Pentium macrocells generally require cooling, which means a fan, which means moving parts, which means that they do not meet selection criteria for telecomunications and military infrastructure, which often requires no moving parts. In general, it's not an issue: most embedded developement is cross-developement, where the compiling is not on the box where the code is expected to run. Even were that not the case, the special-purpose nature of the hardware often means that there is not a real BIOS supporting the hardware function, and it could not boot as a general purpose PC, in any case (it's at least as different as the PC98 -- and usually more different). I think the only place you could make a case is emerging markets in the third world, which are getting leftover equipment from the first world, which translates to them having 386 boxes, and no wayto load FreeBSD, so they go with Linux or some other OS instead. Even that argument is very weak. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
* De: Terry Lambert [EMAIL PROTECTED] [ Data: 2002-12-17 ] [ Subjecte: Re: 80386 out of GENERIC ] Apparently, one of the primary markets for FreeBSD is embedded devices. Are you implying that these people, who are undoubtedly adding and removing lots of things in the kernel, to make things fit, and to make things do their jobs, can't be bothered to use the appropriate CPU settings? -- Juli Mallett [EMAIL PROTECTED] OpenDarwin, Mono, FreeBSD Developer. ircd-hybrid Developer, EFnet addict. FreeBSD on MIPS-Anything on FreeBSD. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Juli Mallett wrote: Are you implying that these people, who are undoubtedly adding and removing lots of things in the kernel, to make things fit, and to make things do their jobs, can't be bothered to use the appropriate CPU settings? Not sure where you got that from Terry's post, but... As a sometimes embedded developer (who also runs FreeBSD on his comparatively screaming Athlon desktop box), being able to run FreeBSD, fresh off a CD, on a quirky 386 embedded toaster and have it run perfectly would be a dream. Of course, that's never been the case. As others have mentioned, you're lucky if you have a working BIOS. There's usually no room for luxuries like a robust device probing system, a nice, standard PCI bus, queriable hardware, etc. Most of your devices are sitting right on the processor bus (and hopefully you've thrown in enough wait states, but if the thing doesn't respond, spin a bit and hammer it with the request again). As long as it's feasible to compile a kernel for a 386, that's all I could ever home for. Just don't go rewriting the scheduler in assembly and use MMX/SIMD instructions... To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
* De: David Cuthbert [EMAIL PROTECTED] [ Data: 2002-12-17 ] [ Subjecte: Re: 80386 out of GENERIC ] Juli Mallett wrote: Are you implying that these people, who are undoubtedly adding and removing lots of things in the kernel, to make things fit, and to make things do their jobs, can't be bothered to use the appropriate CPU settings? Not sure where you got that from Terry's post, but... I misread. As a sometimes embedded developer (who also runs FreeBSD on his comparatively screaming Athlon desktop box), being able to run FreeBSD, fresh off a CD, on a quirky 386 embedded toaster and have it run perfectly would be a dream. Of course, that's never been the case. As others have mentioned, you're lucky if you have a working BIOS. There's usually no room for luxuries like a robust device probing system, a nice, standard PCI bus, queriable hardware, etc. Most of your devices are sitting right on the processor bus (and hopefully you've thrown in enough wait states, but if the thing doesn't respond, spin a bit and hammer it with the request again). As long as it's feasible to compile a kernel for a 386, that's all I could ever home for. Just don't go rewriting the scheduler in assembly and use MMX/SIMD instructions... I don't think anyone wants that to happen (though I wouldn't put it past some people to want to do that). -- Juli Mallett [EMAIL PROTECTED] OpenDarwin, Mono, FreeBSD Developer. ircd-hybrid Developer, EFnet addict. FreeBSD on MIPS-Anything on FreeBSD. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Tue, 17 Dec 2002, Ruslan Ermilov wrote:
On Mon, Dec 16, 2002 at 09:05:40AM +1030, Greg 'groggy' Lehey wrote:
I suppose it would be a good idea to include an alternatvie i386
kernel on the CD-ROM. There may be a space issue, of course. How
many people participating in this thread have an i386 with at least 12
MB of memory and intended to try 5.0 on it? How many of those don't
have a machine to bootstrap off?
Having only alternative i386 kernel is not enough while userland
stuff is still compiled for i486.
Er, userland stuff is still compiled for original i386's, modulo bugs.
E.g., in the i386 endian.h:
% #if defined(_KERNEL) (defined(I486_CPU) || defined(I586_CPU) ||
defined(I686_CPU)) !defined(I386_CPU)
^^^
%
% #define __byte_swap_int(x) \
% __extension__ ({ register __uint32_t __X = (x); \
%__asm (bswap %0 : +r (__X)); \
%__X; })
% #else
%
% #define __byte_swap_int(x) \
% __extension__ ({ register __uint32_t __X = (x); \
%__asm (xchgb %h0, %b0\n\trorl $16, %0\n\txchgb %h0, %b0 \
% : +q (__X)); \
%__X; })
% #endif
The _KERNEL part of the ifdef limits the use of the i486 bswap
instruction to the kernel, so userland is properly pessimized to support
all x86's. The other parts of the ifdef properly pessimize modules to
support all x86's (options don't apply to modules so none of the XXX_CPU's
is defined).
So kernels get the full epsilon of optimizations from turning off i386
support, while userland doesn't get any (not counting ones from optimizing
for non-i386 without breaking i386).
Bruce
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
But still, would it be impossible to have both a GENERIC and a GENERIC386 kernel in the distribution? Or is the whole system compiled in non-386 mode? Even so, if just one site. www.386.freebsd.org were having a 386-enabled version available, wouldn't that make everybody happy? Leif To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: Re[2]: 80386 out of GENERIC
At 4:45 PM -0800 2002/12/15, Avleen Vig wrote: How difficult would the following be to develop, in your opinion? A boot disk image (like the sets of images on the website tm) that will boot on 386's as well as more modern CPU's that can newfs and disklabel your drives, download the source, and let you compile from that point. How many uninterrupted days/weeks would you be willing to allow a make world to run? something to compile with - which can be downloaded with the source all precompile so the do work on all x86 CPU's. If you're compiling the code (as would be necessary, since 386-compatible code would not be included in any of the regular binaries), then there is no such thing as pre-compiled anything. Morever, the concept of compiling something is mutually exclusive to getting a pre-compiled copy of that same something. -- Brad Knowles, [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. -Benjamin Franklin, Historical Review of Pennsylvania. GCS/IT d+(-) s:+(++): a C++(+++)$ UMBSHI$ P+++ L+ !E-(---) W+++(--) N+ !w--- O- M++ V PS++(+++) PE- Y+(++) PGP+++ t+(+++) 5++(+++) X++(+++) R+(+++) tv+(+++) b+() DI+() D+(++) G+() e++ h--- r---(+++)* z(+++) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
RE: Re[2]: 80386 out of GENERIC
On 14-Dec-2002 Alex wrote: Dear/Beste phk, Saturday, December 14, 2002, 10:14:20 PM, you wrote: In message [EMAIL PROTECTED], Alex writes: Dear/Beste Johnson, I read this on the advocacy list. Tuesday, November 19, 2002, 7:56:44 PM, you wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. No it doesn't mean that. FreeBSD still runs on all 386 family CPUs, the only difference is that if you want to run it on a 80386 you need to enable an option in your kernel config file. It will out of the box run on 486 and anything later. It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. My assumption was for such case. Lets say a home user with a pentium II with Windows 98 and his old 386 as a FreeBSD server. 5.0 will be too slow to be useful on an 80386. Better off running 4.x or 2.2.x on a real 80386. Seriously. Also, if/when KSE fully matures the new thread library will need to use special instructions that are only present on the 486 and above (the same instructions (cmpxchg) that are the reason GENERIC doesn't include 80386 by default anymore). Emulating those instructions on a 80386 is expensive and requires kernel privilege, thus someone would have to add support to the kernel to trap the illegal instruction fault on a 80386 and then emulate the instruction. Thus, it's going to be _really_ slow on a 80386. FreeBSD 2.2.x and 4.x are stable and tested OS platforms and I think that for an 80386 they are a much better platform than 5.x or other future versions of FreeBSD will ever be. -- John Baldwin [EMAIL PROTECTED]http://www.FreeBSD.org/~jhb/ Power Users Use the Power to Serve! - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On 15-Dec-2002 Terry Lambert wrote: Alex wrote: It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Yes, this is true. Several of us were annoyed by the change, which appeared at the time to have been done solely to handle the fact that the newly installed device /dev/random sucked too much CPU time to work on a 386. The /dev/random code has since improved to not suck so much CPU time, but the 386 code was not reenabled. The best answer out there is the majority has spoken, with the idea being that if you are deploying on 386 hardware, you are an embedded systems vendor, and are willing to live with the process effectively being a cross-compilation. This has nothing to do with /dev/random. Please stop with the constant FUDing Terry. -- John Baldwin [EMAIL PROTECTED]http://www.FreeBSD.org/~jhb/ Power Users Use the Power to Serve! - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On 15-Dec-2002 Nate Lawson wrote: On Sun, 15 Dec 2002, Erik Trulsson wrote: The only remotely good reason I have heard for removing support for 386 in the default configuration is that having it in would pessimize performance too much for more modern CPUs. How valid that reason is I cannot judge, but I guess it is possible. Could someone enlighten me as to why we don't leave 386 support in for the boot kernel so the floppies will at least boot? Note that performance shouldn't be an issue when installing. After you reboot after you install the kernel you just installed won't boot. Pretty pointless. -- John Baldwin [EMAIL PROTECTED]http://www.FreeBSD.org/~jhb/ Power Users Use the Power to Serve! - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Saturday 14 December 2002 08:53 pm, Terry Lambert wrote: The best answer out there is the majority has spoken, with the idea being that if you are deploying on 386 hardware, you are an embedded systems vendor, and are willing to live with the process effectively being a cross-compilation. Okay, here's a compromise solution for all those people still needing 386 support out of the box: make a 5.0-mini-386.iso image. p.s. I somehow suspect that embedded systems vendors aren't installing from the CDROM. David To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Yes, and then make 5.0-useless-Tandy1000.iso for the other 8 guys that could use it. -Craig - Original Message - From: Johnson David [EMAIL PROTECTED] To: Terry Lambert [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Monday, December 16, 2002 12:55 Subject: Re: 80386 out of GENERIC On Saturday 14 December 2002 08:53 pm, Terry Lambert wrote: The best answer out there is the majority has spoken, with the idea being that if you are deploying on 386 hardware, you are an embedded systems vendor, and are willing to live with the process effectively being a cross-compilation. Okay, here's a compromise solution for all those people still needing 386 support out of the box: make a 5.0-mini-386.iso image. p.s. I somehow suspect that embedded systems vendors aren't installing from the CDROM. David To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Mon, Dec 16, 2002 at 09:55:14AM -0800, Johnson David said: Okay, here's a compromise solution for all those people still needing 386 support out of the box: make a 5.0-mini-386.iso image. p.s. I somehow suspect that embedded systems vendors aren't installing from the CDROM. why is this an issue? 1) supporting every computer made since 1964 is NetBSD's job, not FreeBSD's. 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. 3) if you feel compelled to run old hardware, why not shell out $30 for a 486 system? for $50 you can get a Pentium 166. :-) I'm really keen to see FreeBSD move *forward*. chris --- Chris Doherty chris [at] randomcamel.net I think, said Christopher Robin, that we ought to eat all our provisions now, so we won't have so much to carry. -- A. A. Milne --- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
In message: [EMAIL PROTECTED] Greg 'groggy' Lehey [EMAIL PROTECTED] writes: : I suppose it would be a good idea to include an alternatvie i386 : kernel on the CD-ROM. There may be a space issue, of course. How : many people participating in this thread have an i386 with at least 12 : MB of memory and intended to try 5.0 on it? How many of those don't : have a machine to bootstrap off? 12MB? The last time I tried on a 16MB machine, it core dumped because it ran out of memory. I had to put 24MB in the machine before it would work (I couldn't try 20MB due to onhand SIMMs). Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Monday, 16 December 2002 at 12:02:10 -0700, M. Warner Losh wrote: In message: [EMAIL PROTECTED] Greg 'groggy' Lehey [EMAIL PROTECTED] writes: I suppose it would be a good idea to include an alternatvie i386 kernel on the CD-ROM. There may be a space issue, of course. How many people participating in this thread have an i386 with at least 12 MB of memory and intended to try 5.0 on it? How many of those don't have a machine to bootstrap off? 12MB? Last known safe value. The last time I tried on a 16MB machine, it core dumped because it ran out of memory. I had to put 24MB in the machine before it would work (I couldn't try 20MB due to onhand SIMMs). Ah, good info. OK, how many 24 MB 80386 machines have you (pl) ever seen? How many do you have? Greg -- 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: 80386 out of GENERIC
In message [EMAIL PROTECTED], Johnson David writes: On Saturday 14 December 2002 08:53 pm, Terry Lambert wrote: The best answer out there is the majority has spoken, with the idea being that if you are deploying on 386 hardware, you are an embedded systems vendor, and are willing to live with the process effectively being a cross-compilation. Okay, here's a compromise solution for all those people still needing 386 support out of the box: make a 5.0-mini-386.iso image. Sure, send us the necessary patches to do so and we'll look at it. Poul-Henning PS: If you are asking _us_ to spend _our_ time doing this: Forget it. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
5.0 performance (was: 80386 out of GENERIC)
On Monday, 16 December 2002 at 10:09:48 -0800, Chris Doherty wrote: 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. I'm running it diskless on a K6/233. I'm surprised how snappy it is. Greg -- 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: 5.0 performance (was: 80386 out of GENERIC)
On 2002-12-17 10:57, Greg 'groggy' Lehey [EMAIL PROTECTED] wrote: On Monday, 16 December 2002 at 10:09:48 -0800, Chris Doherty wrote: 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. I'm running it diskless on a K6/233. I'm surprised how snappy it is. I still have the Pentium 133 with 64 MB or memory that I used to run 5.0-CURRENT until a few weeks ago. I haven't got any real numbers, but the general `feel' of the system was pretty good. Trying to build world kernel on a 386 though... now that's a very different story! :) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 5.0 performance (was: 80386 out of GENERIC)
Read the top of /usr/src/UPDATING Explains most of the slow problems. At 03:45 AM 12/17/2002 +0200, Giorgos Keramidas wrote: On 2002-12-17 10:57, Greg 'groggy' Lehey [EMAIL PROTECTED] wrote: On Monday, 16 December 2002 at 10:09:48 -0800, Chris Doherty wrote: 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. I'm running it diskless on a K6/233. I'm surprised how snappy it is. I still have the Pentium 133 with 64 MB or memory that I used to run 5.0-CURRENT until a few weeks ago. I haven't got any real numbers, but the general `feel' of the system was pretty good. Trying to build world kernel on a 386 though... now that's a very different story! :) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
I have a 486 running as a firewall. I honestly didn't even think about trying 5.0 on that thing, forget a 386. So you don't get upset when Windows XP requires a Pentium two million with a bejigabyte of RAM but you complain when 5.0 Won't run on a 386? Adam In message: [EMAIL PROTECTED] [EMAIL PROTECTED] writes: : I don't feel this is a good decision. (I still have a 486, act as a : small server and a 286 witch is in storage) This basally means that : any one who doesn't have the latest processor can't install FreeBSD. : : No it doesn't mean that. : : FreeBSD still runs on all 386 family CPUs, the only difference is that : if you want to run it on a 80386 you need to enable an option in : your kernel config file. : : It will out of the box run on 486 and anything later. One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message -- Adam Migus Network Associates Laboratories (http://www.nailabs.com) TrustedBSD (http://www.trustedbsd.org) | The Power To Secure To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 5.0 performance (was: 80386 out of GENERIC)
On Tue, Dec 17, 2002 at 03:45:22AM +0200, Giorgos Keramidas wrote: On 2002-12-17 10:57, Greg 'groggy' Lehey [EMAIL PROTECTED] wrote: On Monday, 16 December 2002 at 10:09:48 -0800, Chris Doherty wrote: 2) I'm scared that 5.0 is going to be unpleasantly slow on my p2-366, let alone a 386. I'm running it diskless on a K6/233. I'm surprised how snappy it is. I still have the Pentium 133 with 64 MB or memory that I used to run 5.0-CURRENT until a few weeks ago. I haven't got any real numbers, but the general `feel' of the system was pretty good. Trying to build world kernel on a 386 though... now that's a very different story! :) Yup. But the slowness people are noting in general is explained in UPDATING, and is quite understandable at this point in 5.0's evolution. It certainly takes a *lot* longer than 4.7 (test machine 1GHZ Pentium III, 512MB memory, SCSI disk). Also didn't someone mention that GCC has got slower anyway ? -- Regards Cliff Sarginson The Netherlands [ This mail has been checked as virus-free ] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
12MB? The last time I tried on a 16MB machine, it core dumped because it ran out of memory. I had to put 24MB in the machine before it would work (I couldn't try 20MB due to onhand SIMMs). Uhh, I think we should move forward, like everyone else says. I mean, I don't throw computers away, but all my 486's have died now, RIP... Having said that, to the best of my knowledge, there's still a scavenged P-90 with two 3Com NIC's bridging two LANs and acting as an internal router between them after two companies (one of which I used to work at) merged 5 years ago, and they wanted to access the private Frame Relay from their own LAN. It's running FreeBSD 2.2.x on an 80MB HDD with no swap (it was a *very* tight squeeze), and got upgraded to 5MB of RAM (it was installed with 4MB) after the guy running it decided to enable NIS and found it wouldn't work without a little more memory... Heh. I guess that's progress for ya. :) Cheers, AS msg48889/pgp0.pgp Description: PGP signature
Re: 5.0 performance (was: 80386 out of GENERIC)
At 5:58 AM +0100 12/17/02, Cliff Sarginson wrote: Also didn't someone mention that GCC has got slower anyway ? gcc is slower at compiling things. This is very noticeable when you're doing a buildworld. The code which gcc 3.2.1 produces does not seem any slower than the code produced by gcc 2.95.4 (the version in freebsd-stable). -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 5.0 performance (was: 80386 out of GENERIC)
On Tuesday 17 December 2002 12:19 am, Garance A Drosihn wrote: At 5:58 AM +0100 12/17/02, Cliff Sarginson wrote: Also didn't someone mention that GCC has got slower anyway ? gcc is slower at compiling things. This is very noticeable when you're doing a buildworld. The code which gcc 3.2.1 produces does not seem any slower than the code produced by gcc 2.95.4 (the version in freebsd-stable). Actually, in my benchmarks here, the same code tends to yield much faster executables under gcc3, particularly in C++. But these are limited benchmarks (primarily of KDE and my own applications). I'm willing to trade some time on the compile (which, with any luck, happens only once) in exchange for speed in the application (which I may use every day)! :-) -Cliff L. Biffle To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Mon, Dec 16, 2002 at 09:05:40AM +1030, Greg 'groggy' Lehey wrote: On Saturday, 14 December 2002 at 20:53:05 -0800, Terry Lambert wrote: Alex wrote: It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Yes, this is true. Several of us were annoyed by the change, which appeared at the time to have been done solely to handle the fact that the newly installed device /dev/random sucked too much CPU time to work on a 386. That's an interesting apparition. In fact, it was done because the locking primitives for i386 are so different from those for later machines that they would significantly slow down all i[3]86 kernels. Since that's the vast majority, it doesn't make sense. I suppose it would be a good idea to include an alternatvie i386 kernel on the CD-ROM. There may be a space issue, of course. How many people participating in this thread have an i386 with at least 12 MB of memory and intended to try 5.0 on it? How many of those don't have a machine to bootstrap off? Having only alternative i386 kernel is not enough while userland stuff is still compiled for i486. Cheers, -- Ruslan Ermilov Sysadmin and DBA, [EMAIL PROTECTED] Sunbay Software AG, [EMAIL PROTECTED] FreeBSD committer, +380.652.512.251Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age msg48894/pgp0.pgp Description: PGP signature
Re: 80386 out of GENERIC
On Sun, Dec 15, 2002 at 12:18:21AM -0500, Craig Reyenga wrote: Sorry for butting in, but my $.02 is that 386's are old enough that FreeBSD, or any other OS for that matter, shouldn't wait up for them. Why not? An OS in itself should not require a lot of CPU power. They've gotten to the point where they are basically useless except for running older software, which was likely written for them anyways. They are not useless, and if new software has problems running on them it is mostly because a lot of new software is big and bloated without any good reason except for lazy/incompetent programmers. If I had a 386 that I wanted FreeBSD on, I'd crack open the old FreeBSD 3.5 install CD's, assuming it even had a cdrom drive. I understand why people care about supporting older hardware. Reasons such as cost, and the ability to allow code bloat to _really_ manifest itself come to mind. However, a 386 is just too old for words and should be running older software with less features. Less features and more security problems. Considering that security fixes normally don't get applied to the 3.x branch any longer one might want to be a bit careful running that on a computer connected to the Net. Eventually I assume that 4. will be similarily abandoned which means that you will have to run 5.x to have a secure system. Personally I strongly disagree with the notion that hardware that is a mere 10 years old (like some '386s) should be considered too old for words. The only remotely good reason I have heard for removing support for 386 in the default configuration is that having it in would pessimize performance too much for more modern CPUs. How valid that reason is I cannot judge, but I guess it is possible. (And just FYI, my 386 box is happily running 4.7-stable at the moment without any problems and I will probably consider updating it to 5.x when security fixes are no longer automatically applied to 4.x.) -Craig - Original Message - From: Terry Lambert [EMAIL PROTECTED] To: M. Warner Losh [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Saturday, December 14, 2002 23:55 Subject: Re: 80386 out of GENERIC M. Warner Losh wrote: One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Add to this that Bosko's workaround for the CPU bug with PSE/PGE includes loading the kernel at 4M rather than 1M. -- Insert your favourite quote here. Erik Trulsson [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Sun, 15 Dec 2002, Erik Trulsson wrote: The only remotely good reason I have heard for removing support for 386 in the default configuration is that having it in would pessimize performance too much for more modern CPUs. How valid that reason is I cannot judge, but I guess it is possible. Could someone enlighten me as to why we don't leave 386 support in for the boot kernel so the floppies will at least boot? Note that performance shouldn't be an issue when installing. -Nate To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
In message [EMAIL PROTECTED], Nate Lawson wri tes: On Sun, 15 Dec 2002, Erik Trulsson wrote: The only remotely good reason I have heard for removing support for 386 in the default configuration is that having it in would pessimize performance too much for more modern CPUs. How valid that reason is I cannot judge, but I guess it is possible. Could someone enlighten me as to why we don't leave 386 support in for the boot kernel so the floppies will at least boot? Note that performance shouldn't be an issue when installing. Because few if any 80386 computers have the ram it takes to run sysinstall. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
[EMAIL PROTECTED] wrote: Because few if any 80386 computers have the ram it takes to run sysinstall. Was sysinstall around when 386 was new? Just curious what's changed since then to make it bigger. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Was sysinstall around when 386 was new? No, and neither was FreeBSD. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
In message [EMAIL PROTECTED], walt writes: [EMAIL PROTECTED] wrote: Because few if any 80386 computers have the ram it takes to run sysinstall. Was sysinstall around when 386 was new? Just curious what's changed since then to make it bigger. sysinstall arrived in the 486 days. Lots of junk has been added since, but I think most of the bloating is from added kernel stuff, IPv6, PCcard, PCI, USB and so on. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
walt wrote: [EMAIL PROTECTED] wrote: Because few if any 80386 computers have the ram it takes to run sysinstall. Was sysinstall around when 386 was new? Just curious what's changed since then to make it bigger. The sheer number of new drivers, for one thing. -- Daniel C. Sobral(8-DCS) [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] You are as old as you feel. Then I broke a few medical records. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
On Saturday, 14 December 2002 at 20:55:05 -0800, Terry Lambert wrote: M. Warner Losh wrote: One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Add to this that Bosko's workaround for the CPU bug with PSE/PGE includes loading the kernel at 4M rather than 1M. I'm not sure I understand you. i386's have a 32 bit address space, and long ago we loaded at 0xf000 (3.75M). Then we dropped it to 0xc000 (3M). 4M is the end of the address space. Are you talking about something else? Greg -- 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: 80386 out of GENERIC
On Saturday, 14 December 2002 at 20:53:05 -0800, Terry Lambert wrote: Alex wrote: It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Yes, this is true. Several of us were annoyed by the change, which appeared at the time to have been done solely to handle the fact that the newly installed device /dev/random sucked too much CPU time to work on a 386. That's an interesting apparition. In fact, it was done because the locking primitives for i386 are so different from those for later machines that they would significantly slow down all i[3]86 kernels. Since that's the vast majority, it doesn't make sense. I suppose it would be a good idea to include an alternatvie i386 kernel on the CD-ROM. There may be a space issue, of course. How many people participating in this thread have an i386 with at least 12 MB of memory and intended to try 5.0 on it? How many of those don't have a machine to bootstrap off? Greg -- 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: Re[2]: 80386 out of GENERIC
On Sat, 14 Dec 2002, [EMAIL PROTECTED] wrote: FreeBSD still runs on all 386 family CPUs, the only difference is that if you want to run it on a 80386 you need to enable an option in your kernel config file. It will out of the box run on 486 and anything later. It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Too bad. Harsh, but understandable. How difficult would the following be to develop, in your opinion? A boot disk image (like the sets of images on the website tm) that will boot on 386's as well as more modern CPU's that can newfs and disklabel your drives, download the source, and let you compile from that point. That way you don't need to transfer from one box to another, and you get the compiled install you want right away. I'd maybe like ot help with something like this but my abilities and experience are somewhat limited :-) It shouldn't really be that hard should it? You just need the boot disks, with utilities that are: newfs disklabel ifconfig something to download with (fetch? wget?) something to compile with - which can be downloaded with the source all precompile so the do work on all x86 CPU's. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Greg 'groggy' Lehey wrote: Add to this that Bosko's workaround for the CPU bug with PSE/PGE includes loading the kernel at 4M rather than 1M. I'm not sure I understand you. i386's have a 32 bit address space, and long ago we loaded at 0xf000 (3.75M). Then we dropped it to 0xc000 (3M). 4M is the end of the address space. Are you talking about something else? The load is at physical 4M. It avoids the 4K/4M page switch, and has two side effects which make it work without DISABLE_PSE et al. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
80386 out of GENERIC
Dear/Beste Johnson, I read this on the advocacy list. Tuesday, November 19, 2002, 7:56:44 PM, you wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. (I could be wrong) Having a option during the installation to use GENERIC or PENTIUM (IV) would be far better. A option to enable/disable hardware compile one own kernel would be great to. I never owned any SCSI devices in a decade. -- Best regards/Met vriendelijke groet, Alex To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
In message [EMAIL PROTECTED], Alex writes: Dear/Beste Johnson, I read this on the advocacy list. Tuesday, November 19, 2002, 7:56:44 PM, you wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. No it doesn't mean that. FreeBSD still runs on all 386 family CPUs, the only difference is that if you want to run it on a 80386 you need to enable an option in your kernel config file. It will out of the box run on 486 and anything later. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. (I could be wrong) Having a option during the installation to use GENERIC or PENTIUM (IV) would be far better. A option to enable/disable hardware compile one own kernel would be great to. I never owned any SCSI devices in a decade. Be careful that you are not asking for something that is too one-sided. FreeBSD-CURRENT is aimed at the bleeding edge. The fact that it runs AT ALL on your hardware is GOOD. _However_, FreeBSD is not going to let CURRENT be held back to some low denominator that is no longer commercially available. Support has only been dropped for the _default_ install. You can still build a custom kernel for 80386. You are in the minority. You have a way to run FreeBSD-CURRENT, you just don't have it in the default install. :-) M -- Mark Murray Beware! I'm umop ap!sdn To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re[2]: 80386 out of GENERIC
Dear/Beste phk, Saturday, December 14, 2002, 10:14:20 PM, you wrote: In message [EMAIL PROTECTED], Alex writes: Dear/Beste Johnson, I read this on the advocacy list. Tuesday, November 19, 2002, 7:56:44 PM, you wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. No it doesn't mean that. FreeBSD still runs on all 386 family CPUs, the only difference is that if you want to run it on a 80386 you need to enable an option in your kernel config file. It will out of the box run on 486 and anything later. It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. My assumption was for such case. Lets say a home user with a pentium II with Windows 98 and his old 386 as a FreeBSD server. -- Best regards/Met vriendelijke groet, Alex To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: Re[2]: 80386 out of GENERIC
In message [EMAIL PROTECTED], Alex writes: I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. No it doesn't mean that. FreeBSD still runs on all 386 family CPUs, the only difference is that if you want to run it on a 80386 you need to enable an option in your kernel config file. It will out of the box run on 486 and anything later. It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Too bad. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re[3]: 80386 out of GENERIC
It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Another option is to install 4.x and upgrade to 5.x -- Best regards/Met vriendelijke groet, Alex To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re[2]: 80386 out of GENERIC
Dear/Beste Mark, Saturday, December 14, 2002, 11:15:24 PM, you wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. (I could be wrong) Having a option during the installation to use GENERIC or PENTIUM (IV) would be far better. A option to enable/disable hardware compile one own kernel would be great to. I never owned any SCSI devices in a decade. Be careful that you are not asking for something that is too one-sided. I didn't. I asked for something that is many-sided; at leas for as far processors are concerned. (Its a point of view) FreeBSD-CURRENT is aimed at the bleeding edge. The fact that it runs AT ALL on your hardware is GOOD. _However_, FreeBSD is not going to let CURRENT be held back to some low denominator that is no longer commercially available. I got the impressions that it was here to stay. If it comes back when current becomes stable, then you should consider my opinion as not spoken out. You are in the minority. You have a way to run FreeBSD-CURRENT, you just don't have it in the default install. :-) These though really aren't for me. My most powerful PC is a Pentium 133 MHz. Its not impossible to get around. -- Best regards/Met vriendelijke groet, Alex To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
In message: [EMAIL PROTECTED] [EMAIL PROTECTED] writes: : I don't feel this is a good decision. (I still have a 486, act as a : small server and a 286 witch is in storage) This basally means that : any one who doesn't have the latest processor can't install FreeBSD. : : No it doesn't mean that. : : FreeBSD still runs on all 386 family CPUs, the only difference is that : if you want to run it on a 80386 you need to enable an option in : your kernel config file. : : It will out of the box run on 486 and anything later. One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Alex wrote: Support for the 80386 processor has been removed from the GENERIC kernel. The default FreeBSD kernel is now more optimized for modern CPUs. No longer do you have to settle for performance draining compatibility with hardware you haven't owned in a decade. I don't feel this is a good decision. (I still have a 486, act as a small server and a 286 witch is in storage) This basally means that any one who doesn't have the latest processor can't install FreeBSD. (I could be wrong) Having a option during the installation to use GENERIC or PENTIUM (IV) would be far better. A option to enable/disable hardware compile one own kernel would be great to. I never owned any SCSI devices in a decade. Since people here seems a bit touchy on this subject, I'll try to explain it to you. Today's world is one of plugplay devices (of the real type, not that ISA backwards-compatible sham), devices that can come and go (like cardbus and usb), of systems that can control fan speed, enter stand by, sleep and hibernate modes then come back. It is a world where 4 Gb RAM suddenly seems like just a slightly heavier lower-end server, where we count the days 'till Tb hard disks, and make do until then with RAID devices. A world of SMP, and highly sophisticated processors. Well, 80386 was the first step in this direction, and there was much that was left out of it, which people realized later would be needed. It got to a point that supporting 386 implied huge performance penalties for the rest of the ia32 line (and SMP was out of question altogether). Not only that, but 386 was by and large based on ISA devices, and *those* are also out of the window as we expect devices to identify themselves, and negotiate the resources they'll be using. Not only 5.0 GENERIC kernel does not have support for 386, but it also won't play nice at all with old ISA devices. It is not a matter of simply supporting them. They were _hampering_ all other stuff. So we still support them. Just not where they'll get in the way of the newer stuff. But, to be entirely honest, what would someone want with 5.0 on a 386? You go take a look at the new features of FreeBSD 5.0-R, and try to find any that will be of any use on such a system. ACPI, devfs, SMPng, new software RAID system, support for crypto hardware... I'm sorry, but a 386 just won't be able to take advantage of those things. There is really little point in installing 5.0 on a 386, unless you are one of those who _need_ to have the latest. And, if that's the case, then what would you be doing with a 386? :-) Even on the embedded systems market the 386 has passed away. So, when we throw all those things together, the benefits of kicking 386 out of GENERIC far surpass the possible backlash of angry/disappointed 386 owners. And FreeBSD has always carried the flag of taking most benefit of current hardware. There are other operating systems out there that take great pleasure in supporting more obscure hardware (and, yes, 386 is obscure nowadays). -- Daniel C. Sobral(8-DCS) [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] Fundamentalist Debianites, core children of the Linuxen sounds like it could come from the Book of Mormon, or Tolkien on a bad day... To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Alex wrote: It means that you can not install FreeBSD on a 386 unless you have a 486+ machine that can compile a new FreeBSD system and have a way to get that version to the 386. Yes, this is true. Several of us were annoyed by the change, which appeared at the time to have been done solely to handle the fact that the newly installed device /dev/random sucked too much CPU time to work on a 386. The /dev/random code has since improved to not suck so much CPU time, but the 386 code was not reenabled. The best answer out there is the majority has spoken, with the idea being that if you are deploying on 386 hardware, you are an embedded systems vendor, and are willing to live with the process effectively being a cross-compilation. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
M. Warner Losh wrote: One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Add to this that Bosko's workaround for the CPU bug with PSE/PGE includes loading the kernel at 4M rather than 1M. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 80386 out of GENERIC
Sorry for butting in, but my $.02 is that 386's are old enough that FreeBSD, or any other OS for that matter, shouldn't wait up for them. They've gotten to the point where they are basically useless except for running older software, which was likely written for them anyways. If I had a 386 that I wanted FreeBSD on, I'd crack open the old FreeBSD 3.5 install CD's, assuming it even had a cdrom drive. I understand why people care about supporting older hardware. Reasons such as cost, and the ability to allow code bloat to _really_ manifest itself come to mind. However, a 386 is just too old for words and should be running older software with less features. -Craig - Original Message - From: Terry Lambert [EMAIL PROTECTED] To: M. Warner Losh [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Saturday, December 14, 2002 23:55 Subject: Re: 80386 out of GENERIC M. Warner Losh wrote: One problem with most 386 boxes is that they have very little memory. sysinstall is a big, bloated pig dog these days that takes more RAM than most 386 boxes have. This is true also for many 486 boxes too. So even if 386 stuff were in the default kernel, you'd likely have other issues in making sysinstall work and have to do custom hacking... Add to this that Bosko's workaround for the CPU bug with PSE/PGE includes loading the kernel at 4M rather than 1M. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
