Date: Tue, 3 Mar 2020 15:33:25 +0100 From: Cait Smith <kit@hypostasis.com> To: Greg V <greg@unrelenting.technology> Cc: x11@freebsd.org Subject: Re: AMDGPU Picasso on Thinkpad X395 Message-ID: <57C14013-5204-492D-886B-D8FDF5D9A654@hypostasis.com> In-Reply-To: <f5af8291-0f8a-4d70-bfdc-ff2a1ddd0d82@localhost> References: <E98C2011-1E9B-4CCF-9FA5-F2CD6AAE84FA@hypostasis.com> <f5af8291-0f8a-4d70-bfdc-ff2a1ddd0d82@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 3 Mar 2020, at 15:14, Greg V <greg@unrelenting.technology> wrote: >=20 >=20 >=20 > Mar 3, 2020 2:26:51 PM Cait Smith : >=20 >> Hi All >>=20 >> I=E2=80=99ve been away from *BSD for a long while but I=E2=80=99m now = throwing FreeBSD at a Lenovo Thinkpad X395 >=20 > Nice. Does suspend/resume work? I've heard that it does on Ryzen = Mobile, but more confirmations would be good. zzz worked the first time I tried it, and came back, but I only had a = couple of consoles open. I need to do more extensive checks once I have = actual sessions running to see what works and doesn=E2=80=99t=20 >=20 >> I have had wifi and X running under a recent CURRENT and I am = currently rebuilding a 12.1-RELEASE (I probably should have partitioned = the drive and installed both for testing) >=20 > ZFS boot environments. Forget anything you've ever heard about actual = (GPT) partitions :) Thank you for that - I=E2=80=99ve not used them before and still tend to = think in terms of slices - it=E2=80=99s been a while. They do look like = they=E2=80=99ll be a better option once I=E2=80=99ve got my head around = them. >=20 >> With CURRENT and drm-devel it picks up the GPU just fine (attached = dmesg.boot) and I=E2=80=99m happy to play there, although I am no kernel = hacker. I also can see the amdgpu/picasso firmware being loaded with = 12.1-RELEASE and the drivers / firmware compiled and installed from = FreeBSDDesktop, and I=E2=80=99m just now rebuilding xorg from ports to = see how that goes. >>=20 >> My question is this. What is my best course for moderate to = reasonable stability and to best provide feedback (or testing)? >>=20 >> I=E2=80=99m considering following CURRENT, following STABLE with a = CURRENT kernel for the drm-devel port, or 12.1 RELEASE with manually = added drm-v5.0. I=E2=80=99m beginning to think that CURRENT is probably = cleanest, though I may want to drop some of the debug and test code so = that I can get some work done :) >=20 > Yeah, I use current with a non-invariants kernel (e.g. = generic-nodebug) everywhere. >=20 :)=20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?57C14013-5204-492D-886B-D8FDF5D9A654>