Thread (22 messages) 22 messages, 4 authors, 2014-11-04

[PATCH 3.17-rc4 v7 0/6] arm: Implement arch_trigger_all_cpu_backtrace

From: Russell King - ARM Linux <hidden>
Date: 2014-11-04 17:19:54
Also in: lkml

On Tue, Nov 04, 2014 at 05:05:25PM +0000, Daniel Thompson wrote:
This week with the help of one of my colleagues (thanks Tixy) I have
been able to run some tests and figure out what it going on on vexpress-a9.

The summary is that on some GICv1 implementations the bit to enable
group 1 interrupts cannot be accessed using secure memory accesses. More
specifically the presence/absence of the EnableGrp1 bit in the secure
version GICD_CTRL register is implementation defined.

My original patches overlooked this and as a result the existing code
will migrate all interrupts to group but then cannot enable delivery of
group 1 interrupts.

I'm planning to respin the code so it will automatically disable FIQ
support when the EnableGrp1 bit is not implemented. This means
vexpress-a9 will not benefit from FIQ support but will also not be
harmed by it.
That's good news, and it should mean that the Versatile Express starts
booting on the nightly boot tests again.  Thanks for looking into it.

-- 
FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up
according to speedtest.net.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help