From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Huth Date: Wed, 22 Jul 2015 11:40:01 +0000 Subject: Re: powerpc/rtas: Introduce rtas_get_sensor_fast() for IRQ handlers Message-Id: <55AF8111.4080905@redhat.com> List-Id: References: <20150722112527.B36E01402C8@ozlabs.org> In-Reply-To: <20150722112527.B36E01402C8@ozlabs.org> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Michael Ellerman , linuxppc-dev@lists.ozlabs.org, nfont@linux.vnet.ibm.com Cc: paulus@samba.org, anton@samba.org, kvm-ppc@vger.kernel.org On 22/07/15 13:25, Michael Ellerman wrote: > On Fri, 2015-17-07 at 10:46:58 UTC, Thomas Huth wrote: >> The EPOW interrupt handler uses rtas_get_sensor(), which in turn >> uses rtas_busy_delay() to wait for RTAS becoming ready in case it >> is necessary. But rtas_busy_delay() is annotated with might_sleep() >> and thus may not be used by interrupts handlers like the EPOW handler! >> This leads to the following BUG when CONFIG_DEBUG_ATOMIC_SLEEP is >> enabled: > > When did we break this? Hi Michael, the bug has been introduced by commit 587f83e8dd50d22bc0c62e32ec49fd31 ("powerpc/pseries: Use rtas_get_sensor in RAS code") which switched the EPOW handler to use rtas_get_sensor() instead of using rtas_call directly. Also have a look at this thread here: http://www.spinics.net/lists/kvm-ppc/msg10768.html Thomas