[PATCH 09/21] Blackfin: fix early_dma_memcpy() handling of busy channels

From: Mike Frysinger
Date: Thu Jul 16 2009 - 20:09:21 EST


The early logic to locate a free DMA channel and then set it up was broken
in a few ways that only manifested itself when we needed to set up more
than 2 on chip SRAM regions (most board defaults setup 1 or 2). First, we
checked the wrong status register (the destination gets updated, not the
source) and second, we did the ssync before rather than after resetting a
DMA config register.

Signed-off-by: Mike Frysinger <vapier@xxxxxxxxxx>
---
arch/blackfin/kernel/bfin_dma_5xx.c | 29 ++++++++++++++---------------
1 files changed, 14 insertions(+), 15 deletions(-)

diff --git a/arch/blackfin/kernel/bfin_dma_5xx.c b/arch/blackfin/kernel/bfin_dma_5xx.c
index e0bf8cc..9f9b828 100644
--- a/arch/blackfin/kernel/bfin_dma_5xx.c
+++ b/arch/blackfin/kernel/bfin_dma_5xx.c
@@ -253,32 +253,31 @@ void __init early_dma_memcpy(void *pdst, const void *psrc, size_t size)
BUG_ON(src % 4);
BUG_ON(size % 4);

- /* Force a sync in case a previous config reset on this channel
- * occurred. This is needed so subsequent writes to DMA registers
- * are not spuriously lost/corrupted.
- */
- __builtin_bfin_ssync();
-
src_ch = 0;
/* Find an avalible memDMA channel */
while (1) {
- if (!src_ch || src_ch == (struct dma_register *)MDMA_S1_NEXT_DESC_PTR) {
- dst_ch = (struct dma_register *)MDMA_D0_NEXT_DESC_PTR;
- src_ch = (struct dma_register *)MDMA_S0_NEXT_DESC_PTR;
- } else {
+ if (src_ch == (struct dma_register *)MDMA_S0_NEXT_DESC_PTR) {
dst_ch = (struct dma_register *)MDMA_D1_NEXT_DESC_PTR;
src_ch = (struct dma_register *)MDMA_S1_NEXT_DESC_PTR;
+ } else {
+ dst_ch = (struct dma_register *)MDMA_D0_NEXT_DESC_PTR;
+ src_ch = (struct dma_register *)MDMA_S0_NEXT_DESC_PTR;
}

- if (!bfin_read16(&src_ch->cfg)) {
+ if (!bfin_read16(&src_ch->cfg))
+ break;
+ else if (bfin_read16(&dst_ch->irq_status) & DMA_DONE) {
+ bfin_write16(&src_ch->cfg, 0);
break;
- } else {
- if (bfin_read16(&src_ch->irq_status) & DMA_DONE)
- bfin_write16(&src_ch->cfg, 0);
}
-
}

+ /* Force a sync in case a previous config reset on this channel
+ * occurred. This is needed so subsequent writes to DMA registers
+ * are not spuriously lost/corrupted.
+ */
+ __builtin_bfin_ssync();
+
/* Destination */
bfin_write32(&dst_ch->start_addr, dst);
bfin_write16(&dst_ch->x_count, size >> 2);
--
1.6.3.3

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/