Hacker News Reader

Comment by twoodfin | Hacker News Reader

Unlike its ultimate successor in Windows NT, I feel like the story of the design and evolution of the Win32 kernel & the resulting triumph of Windows 95 as an honest-to-goodness protected memory, preemptive multitasking OS hasn’t really been told.

View on HN

Moreover, the fact that it's actually more like a bare-metal hypervisor that runs DOS VMs seems to be little-known. The protection is the bare minimum necessary to keep them isolated - all hardware, unless specifically virtualised by VxDs, is effectively "passed through".

Yeah I'd like a Showstopper book for Windows 95.

Possibly I'm misunderstanding your post. In what sense did Windows 95 have protected memory? Even DOS programs could overwrite important Windows addresses, to say nothing of the convenient win32 APIs MS provided to scribble all over other applications memory space.

Windows 95 absolutely had protected memory: Every 32-bit process had its own address space without direct access to any other process’ memory.

DOS programs were run in V86 mode.

Yeah it was "less unprotected" at best. It was also trivial to deadlock the entire machine with user-mode code (the DirectX system-wide mutex being the most obvious example)

Some books I own that cover the software, and its progression through Windows/386 -> 3.0 -> 3.11 -> 95. They don't go behind the scenes (ie they're not Showstopper!) -

   Pietrek - Windows Internals (1993)
   Schulman, Maxie, Pietrek - Undocumented Windows (1992)
   Pietrek - Windows 95 System Programming Secrets (1995)
   Schulman - Unauthorized Windows 95 (1994)
   Oney - Systems Programming for Windows 95 (Microsoft Press, 1996)
   Williams - DOS and Windows Protected Mode (1994)

I picked them up from eBay and swapmeets for $notmuch. Some are in the Internet Archive's library (where you need a free account, and check books out by the hour)

Reply on news.ycombinator.com