You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've recently been chasing down a regression in newer versions of QEMU booting a WFW 3.11 image, and after analysing the JIT traces I'm fairly sure that it has exactly the same problem that patcher9x solves (which is nicely described in https://www.os2museum.com/wp/those-win9x-crashes-on-fast-machines/).
What has happened is that newer releases of QEMU (from 7.2 onwards) have an optimised LOOP implementation and is now subject to this exact same bug. As per the above article I can confirm that using win /n allows WFW 3.11 to boot successfully, which makes it likely that this is the same NDIS issue that affects Win9x.
Below are the traces obtained from the QEMU JIT where the delay loop is initialised and executed:
Is there any support for WFW 3.11 planned? Since patcher9x already has most of the required infrastructure, I believe it should be possible to devise a patch that can work around the NDIS bug in a similar way.
The text was updated successfully, but these errors were encountered:
VirtualBox also has the same problem with WFW3.11 and Ryzen - after installing the PCNet drivers: black screen with blinking cursor after the Win3.11 startup graphic flashes.
Many "Win 3.1x on VirtualBox" guides/youtube videos have comments asking about that as well.
I've recently been chasing down a regression in newer versions of QEMU booting a WFW 3.11 image, and after analysing the JIT traces I'm fairly sure that it has exactly the same problem that patcher9x solves (which is nicely described in https://www.os2museum.com/wp/those-win9x-crashes-on-fast-machines/).
What has happened is that newer releases of QEMU (from 7.2 onwards) have an optimised
LOOP
implementation and is now subject to this exact same bug. As per the above article I can confirm that usingwin /n
allows WFW 3.11 to boot successfully, which makes it likely that this is the same NDIS issue that affects Win9x.Below are the traces obtained from the QEMU JIT where the delay loop is initialised and executed:
and here is the point where we hit the divide by zero error at address 0x802959f3:
Is there any support for WFW 3.11 planned? Since patcher9x already has most of the required infrastructure, I believe it should be possible to devise a patch that can work around the NDIS bug in a similar way.
The text was updated successfully, but these errors were encountered: