Date: Sat, 13 Jun 2015 22:32:44 +0800 From: Julian Elischer <julian@freebsd.org> To: "O'Connor, Daniel" <darius@dons.net.au> Cc: Maksim Yevmenkin <maksim.yevmenkin@gmail.com>, Andriy Gapon <avg@freebsd.org>, "current@freebsd.org" <current@freebsd.org> Subject: Re: obtaining a minidump from panic() called from NMI handler Message-ID: <557C3F0C.70900@freebsd.org> In-Reply-To: <3FCB93E6-40AA-4673-99F1-14CD145FFFAA@dons.net.au> References: <CAFPOs6qsFZKMVzLEDL5X77H6s5LoTjsc4SkMWgR0D_P8RQG4YQ@mail.gmail.com> <557B1905.80307@FreeBSD.org> <CAFPOs6p5yTvdbJXPOKXuagZxj%2Bu-pE3kt5fsCWCpPVj4vktO%2Bg@mail.gmail.com> <557C2D5F.4080005@freebsd.org> <3FCB93E6-40AA-4673-99F1-14CD145FFFAA@dons.net.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On 6/13/15 10:01 PM, O'Connor, Daniel wrote: >> On 13 Jun 2015, at 09:17, Julian Elischer <julian@freebsd.org> wrote: >> >>>> could it be that more than one CPUs get the NMI at the same time? >>> i guess, its possible. is there an easy way to check for that? >> hard code checks in the code so that all except the first do something different. >> (even only as a debug check).. like write to some location and loop... > How about just something to the effect of.. > if (curcpu != cpu0) > while (1) > ; well, use 'pause' in the loop to save power. or even halt maybe.. (whatever the idle loop uses) and write something to an array ( stopped_by_nmi[curcpu-index] ) on entering the code so the base processor can see what processors entered it. (names not correct) > > -- > Daniel O'Connor > "The nice thing about standards is that there > are so many of them to choose from." > -- Andrew Tanenbaum > GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C > > > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?557C3F0C.70900>