#2473 assigned defect

mvme162 has invalid code in start up

Reported by: Eric Williams Owned by: Needs Funding
Priority: high Milestone: Indefinite
Component: bsps Version: 4.11
Severity: blocker Keywords: bsp m68k
Cc: Blocked By:
Blocking:

Description

TRAP #7 generated in code for bsp_start(). This is same issue as ticket #2204, fix needs to be extended to all m68k BSPs.

Change History (6)

comment:1 Changed on Nov 18, 2015 at 10:54:59 PM by Chris Johns

Thanks Eric.

Is this important for 4.11?

comment:2 Changed on Nov 19, 2015 at 12:14:48 AM by Joel Sherrill

The very prompt response from the gcc community is this is expected behaviour when writing to 0 on the target. The -fno-delete-null-pointer-checks option disables the check.
I am happy to have our code built with this option because we should not be writing to 0 unless it is a special case such as setting up the vector table.
I am now wondering about the state of our other BSPs.

Given this was the response on #2204. Should this option be swept into all m68k BSPs? Or should it be added in a place (not sure offhand of the spot) where it will apply to all targets and BSPs?

I am prone to say we need to do this for 4.11. It likely breaks every BSP on any architecture which touches memory at 0.

comment:3 Changed on Nov 19, 2015 at 1:02:10 AM by Chris Johns

I did not add the option and fixed the BSP as per the commit that closes the ticket. I suggest we fix the m68k BSPs.

comment:4 Changed on Nov 19, 2015 at 3:18:39 PM by Joel Sherrill

OK. Is there any easy way to identify the BSPs impacted? It would be nice to have a master list to work through.

comment:5 Changed on Jan 26, 2017 at 7:16:00 AM by Sebastian Huber

Milestone: 4.11.14.11.2

comment:6 Changed on Feb 15, 2017 at 1:37:51 PM by Sebastian Huber

Milestone: 4.11.2Indefinite
Owner: changed from Chris Johns to Needs Funding
Status: newassigned
Note: See TracTickets for help on using tickets.