OSM ツ

[OSM ツ @LegendaryOSM]

Decided I was fed up with my old blog being clunky so I transitioned it to an Astro blog and so far it's been amazing!

Check out the new series page for RetailKit:
https://t.co/4t9uUo8Ovg

[OSM ツ @LegendaryOSM]

@ar453a I can, I was going to leave it up to the original repo owner to decide if they wanted my changes. I've not really followed their coding style per say either.

[OSM ツ @LegendaryOSM]

I also pushed the improvements for the npt walk to my fork of ps5-hen so folks can mess around with it and see what else they can come up with too.

https://t.co/aWMlLZ6An2

[OSM ツ @LegendaryOSM]

For anyone curious about the internals, this is the AMD64 Architecture Programmer's Manual Volume 2: System Programming I've been digging through. Truly a wealth of information for AMD SVM! https://t.co/Q5vKMTZTb5

[OSM ツ @LegendaryOSM]

Made some improvements to the GPU DMA and the logic for the page table walk, very happy with the results 😄

[OSM ツ @LegendaryOSM]

Made some improvements to the GPU DMA and the logic for the page table walk, very happy with the results 😄

[OSM ツ @LegendaryOSM]

@adamfur7 Yeah this is another option I considered rather than making it so the guest kernel can R/W the pages. Both are valid, I just preferred the simplicity of letting the guest have full access to its pages.

[OSM ツ @LegendaryOSM]

I noticed the recent PS5 HEN defeats XOTEXT by just disabling the NPT on all cores. This works, but it comes with a performance tradeoff, without NPT the HV seems to fall back to shadow page tables. Games seem to have 3x slower loads.

[OSM ツ @LegendaryOSM]

My solution is to just walk the nested CR3 and clear XOTEXT and set the RW flags. Gives essentially the same effect as the Bypervisor bug did with the QA flags. I'll make a PR on Cragson's repo once I get some time :D

[OSM ツ @LegendaryOSM]

@arabpixell eh... not really. I made the fork but the SDK is too broken to work with sadly.

[OSM ツ @LegendaryOSM]

@ImJtagModz https://t.co/DFR6boYEJp This branch might actually be the closest to what I have for my 9.00 since that is mono 5.2.0.

[OSM ツ @LegendaryOSM]

If you told me I would be writing .NET Framework in 2026 I would not have believed you 🤣

Yet here I am doing just that for the PS4. You can check out what I have so far if you are interested. 😄

https://t.co/2G59uBNogD

[OSM ツ @LegendaryOSM]

@ImJtagModz These definitely are stale and haven't been touched since 2017 though.

[OSM ツ @LegendaryOSM]

@ImJtagModz I would love to see what you had done! I had tried something like this with the mono source but I didn't think to check the libmonosgen, great find! 😀

[OSM ツ @LegendaryOSM]

@flat_z The PsmIntDevMode is set via a qa flag.

[OSM ツ @LegendaryOSM]

So I guess I should have spent more time looking at some of the PSM stuff.

Turns out you can just make this INI file "/app0/sce_psm.ini" with "mono jit true" in it. That just works to enable JIT as long as PsmIntDevMode is enabled.

[OSM ツ @LegendaryOSM]

Just noticed it does this as well 👀

[OSM ツ @LegendaryOSM]

@razzlelul @master_s9 Yes I just hesitate to call it that since I really mainly use it for my research. :)

[OSM ツ @LegendaryOSM]

@master_s9 It might not be the best idea to run Fusion with GoldHEN it might cause undefined behavior with the fself and fpkg but that debug kit feature I've not noticed or tested its likely showing up as a side effect of the DECI patches.

[OSM ツ @LegendaryOSM]

https://t.co/LP0bROx5Sb

This was tested being called from SceShellCore similar to how the official way does as I traced the code when there was a method to set the clock via official servers.

I dont have a machine that has the RTC problem so I'm interested if it works :)

[OSM ツ @LegendaryOSM]

I had actually forgot I had been working on a POC for folks to test with the RTC. I had some issues because I couldn't find a time authority that worked with the ps4 because of the limited TLS version.

If folks can try the following and let me know if it works.