[Blackfin] arch: fix bug - trap_tests fails to recover on some tests.
http://blackfin.uclinux.org/gf/project/uclinux-dist/tracker/?action=TrackerItemEdit&tracker_item_id=3719
When the CPLBs get a miss, we do:
- find a victim in the HW table
- remove the victim
- find the replacement in the software table
- put it into the HW table.
If we can't find a replacement in the software table, we accidently
leave a duplicate in the HW table. This patch ensures that duplicate
is marked as not valid.
What we should do is find the replacement in the software table, before
we find a victim in the HW table - but its too late in the release cycle
to do that much restructuring of this code.
Rather that duplicate code, connect Hardware Errors (irq5) into trap_c,
so user space processes get killed properly.
The rest of irq_panic() can be moved into traps.c (later)
There is still a small corner case that causes problems when a
pheriperal interrupt goes off a single cycle before a user space
hardware error. This causes a kernel panic, rather than the user
space process being killed.
But, this checkin makes things work in 99.9% of the cases, and is a vast
improvement from what is there today (which fails 100% of the time).
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
diff --git a/arch/blackfin/mach-common/cplbmgr.S b/arch/blackfin/mach-common/cplbmgr.S
index faca1ab..f5cf3ac 100644
--- a/arch/blackfin/mach-common/cplbmgr.S
+++ b/arch/blackfin/mach-common/cplbmgr.S
@@ -190,7 +190,14 @@
[P0 - 4] = R0;
R0 = [P0 - 0x100];
[P0-0x104] = R0;
-.Lie_move:P0+=4;
+.Lie_move:
+ P0+=4;
+
+ /* Clear ICPLB_DATA15, in case we don't find a replacement
+ * otherwise, we would have a duplicate entry, and will crash
+ */
+ R0 = 0;
+ [P0 - 4] = R0;
/* We've made space in the ICPLB table, so that ICPLB15
* is now free to be overwritten. Next, we have to determine
@@ -515,14 +522,23 @@
R0 = [P0++]; /* move data */
[P0 - 8] = R0;
R0 = [P0-0x104] /* move address */
-.Lde_move: [P0-0x108] = R0;
+.Lde_move:
+ [P0-0x108] = R0;
+
+.Lde_moved:
+ NOP;
+
+ /* Clear DCPLB_DATA15, in case we don't find a replacement
+ * otherwise, we would have a duplicate entry, and will crash
+ */
+ R0 = 0;
+ [P0 - 0x4] = R0;
/* We've now made space in DCPLB15 for the new CPLB to be
* installed. The next stage is to locate a CPLB in the
* config table that covers the faulting address.
*/
-.Lde_moved:NOP;
R0 = I0; /* Our faulting address */
P2.L = _dpdt_table;