eipä toi #vastaamo-todistusaineisto jätä hirveästi arvailun varaan. huvin ja harrastuksen vuoksi kiinnostaisi tietää, miten #KRP murti salauksen, mutta se jäänee pimentoon.

@uheplugins

@uheplugins https://tal-software.com/products/tal-dac

This #Behringer release made my realize what was missing from authentic #Roland JP-8000 sound when using @uheplugins #Diva: I need to add 12-bit bit reduction :-) With #ToguAudoLine #DAC added it hits the spot for me. https://www.youtube.com/watch?v=Trv8i3PyAAY&t=568s

@animatek As far as sound cards go I love my RME Babyface Pro FS :-) I boot it in class compliant mode when using Linux and proprietary mode when on macOS.

I need to really in-depth my RISC-V understanding and shift the focus away from the bug I've been struggling with (see various posts) :-) Reading all articles I can find related to CPU state changes, and also CVA6 user manual. So far have survived with trial and error but now I feel that I just don't have understanding of some bigger ideas... I should have taken my focus away from the issue at hand weeks ago but well you know...

@animatek #Pipewire has pretty much made the audio routing better than other operating systems. I use my Mac to do mostly music but when I’m on my work PC, I use Bitwig with Pipewire backend. I have Debian running on that.

With Pipewire I’m able to do stuff that I need Loopback on Mac, i.e. route audio somewhat freely between apps.

3rd party plugins that I actually use work too. For U-he stuff I have native Linux versions and FabFilter goodies I use through #yabridge.

Generally Windows plugins that are programmed with love in the first place (like FF stuff) works super solid with Yabridge, sometimes even more stable than real Windows. For super proprietary stuff like XLN Audio stuff I keep them only on my Mac.

Pipewire has sort of fixed things that I had to complain in the past with Linux audio.

I sort of like constructing tracks on my Linux PC given the more limited set of plugins. It keeps the focus on track construction, not on swapping plugins :-) I tend to finish the tracks on my Mac.

You should probably consider more realtime oriented kernel for the installation. I personally use https://liquorix.net/.

PS. I would not touch Ubuntustudio style distro’s because they tend to tune things sort of “old world” ways and sometimes even counter-productive. You are better off with e.g. standard Debian or Ubuntu installation. In addition to more real-time oriented kernel, you might want to add threadirqs to /etc/default/grub and install and systemctl enable rtirq with its default settings , This will give priority to your USB audio interface. I’m not sure if you even need at modern times to do anything to /etc/security/limits.conf given that by default Pipewire daemons do run already with real-time priorities and Bitwig has a Pipewire backend option.

To summarize, a stock distribution, real-time’ish kernel, rtirq and Pipewire would be me go-to list. At least I experience no glitches. And then the desktop is not tuned too heavily to a single purpose (audio), and configuration sort of stands better time as it is not too specific in its adjustments. Not that different from Windows or macOS in the end. In both you also need to few tweaks here and there to get a good result…

#CVA6 #PDF #manual: https://readthedocs.com/projects/openhw-group-cva6-user-manual/downloads/ #note

@nirbheek Actually also great for baremetal ARM testing since I can compile kernel images for RPi 3B+ without cross-compiler...

@nirbheek Made me pour almost 3k to my 32 GB RAM Mac Mini M2 Pro early this year so I guess it is still a working business model :------) Ya, I'm a bitch. It has been useful tho for native ARM64 (or at least "real" VM instead of emulation) kernel testing and music production.

@starshine password at any frequency pwgen -y 12 1

One constraint good to be aware of is that Keystone enclaves are always contiguous physical regions. Page tables are only tool for maintaining address space.

Have been passive at #LKML because I’m stuck with this page table sync bug in #Keystone: https://github.com/keystone-enclave/keystone/issues/378.

Snippet where it trips:

    sfence.vma
    csrw satp, a0 // switch to virtual addresssing
    sfence.vma

mtval is at the 2nd sfence.vma when misaligned store happens.

This never happens on QEMU and I presume that not on all hardware but I’ve tested this only with CVA6 (running on Genesys2 FPGA board). On QEMU everything just works.

At least, shouldn’t the sfence.vma be sort “useless” as it should continue where stvec is set, right? I guess here one thing to try out would be to simply remove the second sfence.vma?

I’m really just learning RISC-V sync opcodes and do not know e.g. when I should use also fence.i. Just trying different things without tbh knowing exactly what I’m doing yet…

In head.S there is also .align 2 after csrw and I have to admit that I don’t know exactly why… This bug is haunting me…

@vbabka I commented actually along the lines to some thread that it would be nice if there was "bridge" between eBPF and AML :-) It would be useful for sure.

Must for any Windows installation: https://wpd.app/

I mostly use #neovim but I also have paid version of #sublimetext. #GUI editor has its place e.g. when you brainstorm stuff with a group and similar contexts.

For me Sublime Text is less proprietary than #vscode because I get a way with a fixed sum of money. I would find it more proprietary deal to be connected 247 to #microsoft.

Microsoft cannot really ad itself even today to supporting users privacy given that for any #Windows installation you really should run tools such as WPD.

IMHO privacy is the worst currency to pay assets with…

Now I get why there’s been so much #ACPI discussion recently in my feed. IMHO, the whole topic is not even worth of essay because you would need to compare it something that also enables universal management interface for heterogenous data center environment. ACPI is ugly but it is also universal. I’d like to see something less ugly implementing the same role.

When it comes to bloat I would rather focus on desktop applications implemented with #JavaScript and #HTML eating all your computing resources. It is ugly and bloated in a whole new level. Web desktop truly sucks

.,, and biggest contribution #vscode has made to the software ecosystem is turning even the freaking text editor as malware.

@stefanha Thanks! This gave me some pointers. So what comes after machine type is specific to that machine.

The problem here is that bootrom is the first stage bootloader i.e. the whole power on starts from that code and it defines root of trust. I’ll need to study Spike documentation on how to do equivalent thing as with virt.

This helped anyway so that I don’t waste time on something useless. Thank you!

OK, something probably pretty basic with #QEMU I need to ask.

I have this command-line, which is generated by #Keystone #RISCV #TEE build:

/home/jarkkojs/work/keystone/qemu/build/riscv64-softmmu/qemu-system-riscv64 \
   $DEBUG \
   -m 2G \
   -nographic \
   -machine virt,rom=/home/jarkkojs/work/keystone/build/bootrom.build/bootrom.bin \
   -bios /home/jarkkojs/work/keystone/build/sm.build/platform/generic/firmware/fw_payload.bin \
    \
   -netdev user,id=net0,net=192.168.100.1/24,dhcpstart=192.168.100.128,hostfwd=tcp::${HOST_PORT}-:22 \
   -device virtio-net-device,netdev=net0 \
   -device virtio-rng-pci \
   -smp $SMP

QEMU has support for #Spike, which is #CVA6 emulation.

So when I do to the above command-line simply s/virt,/spike,/g, I get this error:

$ build/scripts/run-qemu.sh
**** Running QEMU SSH on port 3617 ****
qemu-system-riscv64: Property 'spike-machine.rom' not found

How do I end up to this error message? Why does it seek a filename like that? Without rom= I can run Spike emulation but I need that bootrom for the application.

@bsdphk @mjg59 One innovation in this area that might be cool or even useful, would be to translate AML to eBPF and use unified interpreter for both inside kernel (perhaps). It would be useful for e.g. ACPI debugging when you could try ACPI table fixup with eBPF program... Not sure how feasible would be but if it was feasible I could see applications...