Thread View: comp.sys.next.sysadmin
3 messages
3 total messages
Started by Barry Lustig
Thu, 21 Jan 1993 19:50
NeXTs running 3.0 *cannot* swap to a raw partition
Author: Barry Lustig
Date: Thu, 21 Jan 1993 19:50
Date: Thu, 21 Jan 1993 19:50
24 lines
1056 bytes
1056 bytes
There has been speculation recently about whether a NeXT can swap to
a raw partition or not. I decided to do a little research on this
point and found out that you cannot. The reason for this should be
clear from the following:
Current directory is /
Reading symbol data from sdmach...
(no debugging symbols found)...done.
(gdb) disassemble swapon
Dump of assembler code for function swapon:
0x4068288 <swapon>: rts
End of assembler dump.
(gdb)
NeXT removed the code for the swapon system call from the kernel.
They would have made life a lot easier for people if they also removed
the call from the system library.
barry
--
Re: NeXTs running 3.0 *probably* cannot swap to a raw partition
Author: Carl Edman
Date: Thu, 21 Jan 1993 19:32
Date: Thu, 21 Jan 1993 19:32
30 lines
2135 bytes
2135 bytes
In article <1993Jan21.195003.6023@ictv.com> barry@ictv.com (Barry Lustig) writes:
There has been speculation recently about whether a NeXT can swap to
a raw partition or not. I decided to do a little research on this
point and found out that you cannot. The reason for this should be
clear from the following:
Current directory is /
Reading symbol data from sdmach...
(no debugging symbols found)...done.
(gdb) disassemble swapon
Dump of assembler code for function swapon:
0x4068288 <swapon>: rts
End of assembler dump.
(gdb)
NeXT removed the code for the swapon system call from the kernel.
They would have made life a lot easier for people if they also removed
the call from the system library.
Not quite so fast. All you found out that the BSD system call to
enable swaping is a dummy call. From that one might conclude that the
NeXT can not swap at all just as quickly as that it can not swap to a
raw partition. Fortunately that is not so as virtual memory is one of
the system services which is provided by Mach through the mach_swapon
system call and not BSD on a NeXT. And mach_swapon is _not_ a dummy
call. Careful analysis of this call and its subroutines might reveal
that indeed it is impossible to swap to a raw partition (as I believe
it is). This analysis is left as an exercise to the reader.
Carl Edman
Re: NeXTs running 3.0 *probably* cannot swap to a raw partition
Author: Barry Lustig
Date: Mon, 25 Jan 1993 07:44
Date: Mon, 25 Jan 1993 07:44
23 lines
1598 bytes
1598 bytes
In article <CEDMAN.93Jan21153237@capitalist.princeton.edu> cedman@princeton.edu (Carl Edman) writes: Not quite so fast. All you found out that the BSD system call to enable swaping is a dummy call. From that one might conclude that the NeXT can not swap at all just as quickly as that it can not swap to a raw partition. Fortunately that is not so as virtual memory is one of the system services which is provided by Mach through the mach_swapon system call and not BSD on a NeXT. And mach_swapon is _not_ a dummy call. Careful analysis of this call and its subroutines might reveal that indeed it is impossible to swap to a raw partition (as I believe it is). This analysis is left as an exercise to the reader. Carl Edman As I mentioned to Carl in a mail message, I did look at mach_swapon as well. As far as I can tell, mach_swapon uses the vnode pager code. This provides for paging to a file, not to a raw device. One can also look at the virtual memory sub-system of 386bsd or Berkeley net-2. Both of them use the mach virtual memory system. barry --
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads