5 ms·

It won't convincingly look like Windows 95 to me without a ton of Invalid Page Faults in kernel32.dll. (Can somebody give me a basic explanation of what that w

by mbg721 5y ago

It won't convincingly look like Windows 95 to me without a ton of Invalid Page Faults in kernel32.dll. (Can somebody give me a basic explanation of what that was??)

ronsor 5y ago
The best guess I have is buggy programs passing invalid pointers to kernel32 functions (e.g. OpenFile/WriteFile).
pleb_nz 5y ago
invalid page faults
dredmorbius 5y ago
Install the BSOD x11-screensaver hack. (I used this for a while in rotation with others ... cow-orkers would occasionally tell my my laptop had crashed. I'd enter my password and continue working...)
toast0 5y ago
Ok, so a page fault is when you try to access a memory address that's not mapped by the MMU, or a write to a memory address that's mapped as read only. The MMU works in pages though, not individual addresses, on x86, standard pages are 4096 bytes, although larger pages are available. A valid page fault is when the OS has allocated that page in the current context, but not actually mapped it; it could be a not yet mapped in page from a memory-mapped file, or a page that was swapped to disk, or a page that was allocated but not yet used (not sure if win95 does that), or a write to a page that is allocated as read-write but is currently mapped as read-only and needs copy-on-write (COW) processing now. An invalid page fault is a read or a write to a page that the MMU doesn't have a mapping for and the OS also doesn't have a mapping for. These errors can be caused by many things; often a programming error leads to pointers being overwritten with unrelated data, when they're dereferenced it could access the wrong data or it could cause an invalid page fault. Hardware errors can also result in bad addresses in a pointer. Sometimes it's unintended interactions between running software or software and hardware.
flomo 5y ago
One thing is PC hardware back then was often complete crap. You could run diagnostics and get a lot of memory and disk errors, but Windows 95 still ran "okay". It was definitely a lot more stable if you spent the money on good stuff. Part of it was low expectations too. The half-life for Windows 3.1 crashes was about 4 hours. If your computer stayed up all day, that was considered a massive improvement. It took people a long time to discover the "49 day bug" that was in both Windows and OS/2.
voidfunc 5y ago
My lasting memory of all Win9x editions will be "The program has performed an illegal operation and will be shut down" error messages which were rampant back then. Also I remember people freaking out about those messages.

by mbg721 5y ago

It won't convincingly look like Windows 95 to me without a ton of Invalid Page Faults in kernel32.dll. (Can somebody give me a basic explanation of what that was??)

ronsor 5y ago

The best guess I have is buggy programs passing invalid pointers to kernel32 functions (e.g. OpenFile/WriteFile).

pleb_nz 5y ago

invalid page faults

dredmorbius 5y ago

Install the BSOD x11-screensaver hack. (I used this for a while in rotation with others ... cow-orkers would occasionally tell my my laptop had crashed. I'd enter my password and continue working...)

toast0 5y ago

Ok, so a page fault is when you try to access a memory address that's not mapped by the MMU, or a write to a memory address that's mapped as read only. The MMU works in pages though, not individual addresses, on x86, standard pages are 4096 bytes, although larger pages are available. A valid page fault is when the OS has allocated that page in the current context, but not actually mapped it; it could be a not yet mapped in page from a memory-mapped file, or a page that was swapped to disk, or a page that was allocated but not yet used (not sure if win95 does that), or a write to a page that is allocated as read-write but is currently mapped as read-only and needs copy-on-write (COW) processing now. An invalid page fault is a read or a write to a page that the MMU doesn't have a mapping for and the OS also doesn't have a mapping for. These errors can be caused by many things; often a programming error leads to pointers being overwritten with unrelated data, when they're dereferenced it could access the wrong data or it could cause an invalid page fault. Hardware errors can also result in bad addresses in a pointer. Sometimes it's unintended interactions between running software or software and hardware.

flomo 5y ago

One thing is PC hardware back then was often complete crap. You could run diagnostics and get a lot of memory and disk errors, but Windows 95 still ran "okay". It was definitely a lot more stable if you spent the money on good stuff. Part of it was low expectations too. The half-life for Windows 3.1 crashes was about 4 hours. If your computer stayed up all day, that was considered a massive improvement. It took people a long time to discover the "49 day bug" that was in both Windows and OS/2.

voidfunc 5y ago

My lasting memory of all Win9x editions will be "The program has performed an illegal operation and will be shut down" error messages which were rampant back then. Also I remember people freaking out about those messages.