<html> <head> <base href="https://bugs.freedesktop.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_REOPENED " title="REOPENED - Radeonsi on Grenada cards (r9 390) exceptionally unstable and poorly performing" href="https://bugs.freedesktop.org/show_bug.cgi?id=91880#c125">Comment # 125</a> on <a class="bz_bug_link bz_status_REOPENED " title="REOPENED - Radeonsi on Grenada cards (r9 390) exceptionally unstable and poorly performing" href="https://bugs.freedesktop.org/show_bug.cgi?id=91880">bug 91880</a> from <a class="email" href="mailto:christoph.seifert27@googlemail.com" title="Christoph Seifert <christoph.seifert27@googlemail.com>"> Christoph Seifert</a> <pre>For me switching power states does also result in a system freeze. With radeon.dpm = 0 everthing is working properly but slowly. If I switch manually to another power profile (e.g. echo high > /sys/class/drm/card0/device/power_profile) I got a freeze likewise. With radeon.dpm = 1 the freeze happens after a few seconds of video playback or some other load, ergo if the card changes its power profile. The work around (k firmware) from <a href="show_bug.cgi?id=91880#c115">comment #115</a> does not work for me. Same behaviour as without. So I tried disabling specific DPM features as suggested by Alex Deucher in <a href="show_bug.cgi?id=91880#c60">comment #60</a>. Disabling mclk (pi->mclk_dpm_key_disabled = 1) does the trick for my card but the performance is similar to the radeon.dpm = 0 kernel parameter. With radeon.dpm = 1 and only mclk disabled, the sclk (core clock) adjusts just fine. High on load, low on no load. The mclk (memory clock) is just fixed at 150 Mhz. So the freezes seems to be caused by switching memory clock. Any hints for digging deeper? My card is a MSI Radeon R9 390 too. Linux Kernel 4.8.6 Mesa 13.0.0rc2</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>