Date: Thu, 11 Feb 2016 13:57:35 +0000 From: bugzilla-noreply@freebsd.org To: freebsd-bugs@FreeBSD.org Subject: [Bug 207068] hwpmc wrap around/sign extension Message-ID: <bug-207068-8-FARpv8XsKI@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-207068-8@https.bugs.freebsd.org/bugzilla/> References: <bug-207068-8@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D207068 --- Comment #7 from Konstantin Belousov <kib@FreeBSD.org> --- (In reply to joss.upton from comment #6) I believe that the behaviour of reading MSR 0xc1 is architectural ? Anyway= , on sandybridge: sandy% sudo cpucontrol -m 0xc1=3D0x80000000 /dev/cpuctl0 sandy% sudo cpucontrol -m 0xc1 /dev/cpuctl0 MSR 0xc1: 0x0000ffff 0x80000000 SB does have full-width PMC write capability: sandy% sudo cpucontrol -m 0x345 /dev/cpuctl0 MSR 0x345: 0x00000000 0x000031c3 bit 13 (FW_WRITE) is set But my concern is that I cannot reproduce the issue with the following scri= pt: for x in $(jot 8); do pmcstat -P CPU_CLK_UNHALTED_CORE perl ./loop.pl 2>/dev/null &; done Unpatched kernel must exhibit the problem because it does not write to PMC through aliases and all writes are clipped. I want to be able to reproduce this before commit, at least I need to validate the change and to understand that it is complete (I agree that this is the right approach, at least). --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-207068-8-FARpv8XsKI>