There's no need to keep it around. DSP should stop trying to access
system memory.
Signed-off-by: Felipe Contreras <felipe.contreras@xxxxxxxxx>
---
drivers/staging/tidspbridge/core/ue_deh.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/tidspbridge/core/ue_deh.c b/drivers/staging/tidspbridge/core/ue_deh.c
index ee2d23a..50868a4 100644
--- a/drivers/staging/tidspbridge/core/ue_deh.c
+++ b/drivers/staging/tidspbridge/core/ue_deh.c
@@ -52,8 +52,6 @@
#include <dspbridge/io_sm.h>
-static void *dummy_va_addr;
-
int bridge_deh_create(struct deh_mgr **ret_deh_mgr,
struct dev_object *hdev_obj)
{
@@ -154,6 +152,7 @@ void bridge_deh_notify(struct deh_mgr *deh_mgr, u32 ulEventMask, u32 dwErrInfo)
.element_size = HW_ELEM_SIZE16BIT,
.mixed_size = HW_MMU_CPUES,
};
+ void *dummy_va_addr;
if (!deh_mgr)
return;
@@ -214,6 +213,9 @@ void bridge_deh_notify(struct deh_mgr *deh_mgr, u32 ulEventMask, u32 dwErrInfo)
HW_MMU_TRANSLATION_FAULT);
dump_dsp_stack(dev_context);
dsp_clk_disable(DSP_CLK_GPT8);
+
+ hw_mmu_disable(resources->dw_dmmu_base);
+ free_page((unsigned long)dummy_va_addr);
break;
#ifdef CONFIG_BRIDGE_NTFY_PWRERR
case DSP_PWRERROR:
@@ -276,6 +278,4 @@ int bridge_deh_get_info(struct deh_mgr *deh_mgr,
void bridge_deh_release_dummy_mem(void)
{
- free_page((unsigned long)dummy_va_addr);
- dummy_va_addr = NULL;
}
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
