hrm.. any one have a VM (doesn't need to be illumos) running on qemu w/ a sata disk?

if i can get the output of sg_sat_identify, I might be able to omit the stuff a bit easier..

bhyve doesn't advertise the GPL (general purpose logging) bit in it's identify info, which means no logs... if qemu does the same, that'd probably be the easiest

hello all, is there any procedure to bind process? I have a question here pastebin.com/Eqf49kVz , anyone could take a look? a big thanks

Hi I have a question.

When booting with uefi, the bootloader obtains the framebuffer from gop and writes physical address to the fb_info structure in gfx_private. console continues to use the framebuffer after boot.

At end of booting, ExitBootServices is called, resolution modifications are no longer possible, but the framebuffer is still accessible.

So in theory x11 could be used with a framebuffer when uefi boot?

Why is there no such attempt?

once I ported x11 framebuffer module to dilos which uses the initialized device

apt.tambov.ru/tde/dilos_14.1.0.png look at the image resolution

Oh... good to know.

After searching through the source code,

I think the gfx_fb.c under fcil folder, can be useful to how to use fb.

[illumos-gate] 16537 cxgbe: update firmware version to 1.27.5.0 -- Dan McDonald <danmcd⊙mi>

Any "linker aliens" lunking here? I'm looking for advice on how to interpose on some libc and libidmap functions in a test program that exercises some libsec functionss (acl_totext). It seems that direct bindings (in libsec) works against me in this effort.

I suspect an audit library may be able to override the direct binding entirely. But it's been a while.

My current test program just recompiles the libsec functsions as an old-fashioned static lib.a which works fine for this, but is kind of ugly.

I'd like to turn this into a usr/src/test/libsec-test but would rather not include the static lib hack in that, so looking for a better way.

Yeah, the la_symbind* interfaces in the audit library I think can basically allow them to be changed somewhat arbitrarily.

It was how I first prototyped umem ptc before doing the genasm.

I just added this to the test program link and it works: -Wl,-zinterpose

Linker magic! :)

That's good because I really like this approach for unit tests, where I control _both_ the entry points into the code under test _and_ any support functions it calls.

In this case I'm faking up some specific names for UIDs, GIDs, SIDs, names, etc.

In this case I'm faking up some specific things for UIDs, GIDs, SIDs, names, etc.

uggggh multiboot2

why

I spy a few familiar names in the credits section of the just published POSIX.1-2024/XPG8 standard today... Jörg Schilling, Ryan Zezeski, Casper Dik...

nice rpz