Re: git: 3e6e4e4a0d42 - main - hwpstate: add CPPC support for pstate driver on AMD
Date: Tue, 03 Feb 2026 21:12:19 UTC
In message <3081716.hHqAuc6tWs@ravel>, Olivier Certner writes: > --nextPart2431391.THHZn3L5Ee > Content-Transfer-Encoding: 7Bit > Content-Type: text/plain; charset="utf-8"; protected-headers="v1" > From: Olivier Certner <olce@freebsd.org> > To: Dag-Erling =?UTF-8?B?U23DuHJncmF2?= <des@freebsd.org> > Date: Tue, 03 Feb 2026 16:14:10 +0100 > Message-ID: <3081716.hHqAuc6tWs@ravel> > In-Reply-To: <86ms1pdgnv.fsf@ltc.des.dev> > MIME-Version: 1.0 > > > This doesn't work. You are reusing the same OID already used by the > > Intel subdriver. The result is that whichever one happens to be first > > on the linker command line works and the other one doesn't. You have to > > either use a different OID for each subdriver or have both subdrivers > > share a single boolean. > > Yes, we are aware. There's already https://reviews.freebsd.org/D54528 about > that. > > We did not feel this was urgent, especially since I intend to retire `machdep > .hwpstate_pkg_ctrl` from the AMD driver entirely (and obsolete it in the Inte > l driver) in favor of an easier to use interface, but this has lasted for a w > hile now and could prevent the same knob from working in existing Intel cases > . > > I'll move D54528 in the coming days. > > -- > Olivier Certner One of the subsequent commits broke the commit that allows disabling of CPPC through a tunable. Setting sysctl machdep.hwpstate_amd_cppc_enable failed: 12 -- Cheers, Cy Schubert <Cy.Schubert@cschubert.com> FreeBSD UNIX: <cy@FreeBSD.org> Web: https://FreeBSD.org NTP: <cy@nwtime.org> Web: https://nwtime.org e**(i*pi)+1=0