I hope to have it running on an 11/05 and a 11/34 late this year'
Paul
On 9/7/07, Jerome H. Fine <jhfinedp3k at compsys.to> wrote:
Just as there were inquiries about using PDP-11 hardware, I thought
that it might be helpful to know if anyone uses RT-11 and for what
purpose.
For myself, I have the addiction of fixing RT-11 bugs
in the operating system as well as making enhancements.
Ten years ago, I even produced most of the Y2K changes
that were required for V05.04G of RT-11 for a customer
who could not wait for Mentec. Eventually, I want to
produce a set of Y9K changes to allow years up to 9999
to be used.
Other enhancements include SL: and a new pseudo device
driver, SB: (SymBolic device list device driver) which
is similar to the VMS SNL (Symbolic Name List), but
operates only on a device which supports a directory
as does PATH for DOS. At one point it was named PH:
for Path Handler, but SB: seems a better choice. Even
more reasonable would be Symbolic device List or SL:, but
SL: is already used by the Single Line editor (or the DOSKEY
interface). Symbolic Debugger is also taken for SD:
Lately, I mostly use Ersatz-11 and I have been making
changes and enhancements to the HD: (Hypothetical Disk)
device driver that John Wilson originally wrote in 1995.
The challenge I set myself was to produce a VM(X).SYS
equivalent that is:
(a) Faster - about 3 times as fast
(b) Higher capacity - a full 65536 blocks vs 8192 blocks
(c) Smaller number of LOADed words in low memory
(d) Support the command SET VM: [NO]WRITE
plus a number of other useful features.
Well, one idea led to another and other versions are also
being tested which support 8 units and eventually 64 units
(with monitors that have that support as well). If there is
any interest, eventually a translation table similar to the
one used by DU(X).SYS is possible. In addition, full 32 bit
block number support will be done eventually which will
allow 2 TeraByte disk drives.
The other HD: related code that is also interesting is the
ability to interface the HD: device directly from user code
WITHOUT a device driver. So while the HD: is about twice
as fast as DU(X).SYS when a device driver HD(X).SYS is used,
the direct user interface (which avoids all the overhead of
a system EMT call) is about twice as fast again. For this
user interface, 32 bit blocks numbers are already possible.
Under a Windows (YEK !! Double YEK) operating system or even
just DOS, all disk access is completed before the HD: returns
control to the RT-11 operating system. Consequently, no
interrupts are required. I am not sure about Linux operation,
but if anyone does know, please advise.
Sincerely yours,
Jerome Fine