struct dma_async_tx_descriptor *tx;
enum dma_ctrl_flags dma_flags = DMA_PREP_PQ_DISABLE_P;
+ if (submit->flags & ASYNC_TX_FENCE)
+ dma_flags |= DMA_PREP_FENCE;
dma_dest[1] = dma_map_page(dev, dest, 0, len, DMA_BIDIRECTIONAL);
dma_src[0] = dma_map_page(dev, srcs[0], 0, len, DMA_TO_DEVICE);
dma_src[1] = dma_map_page(dev, srcs[1], 0, len, DMA_TO_DEVICE);
async_tx_submit(chan, tx, submit);
return tx;
}
+
+ /* could not get a descriptor, unmap and fall through to
+ * the synchronous path
+ */
+ dma_unmap_page(dev, dma_dest[1], len, DMA_BIDIRECTIONAL);
+ dma_unmap_page(dev, dma_src[0], len, DMA_TO_DEVICE);
+ dma_unmap_page(dev, dma_src[1], len, DMA_TO_DEVICE);
}
/* run the operation synchronously */
struct dma_async_tx_descriptor *tx;
enum dma_ctrl_flags dma_flags = DMA_PREP_PQ_DISABLE_P;
+ if (submit->flags & ASYNC_TX_FENCE)
+ dma_flags |= DMA_PREP_FENCE;
dma_dest[1] = dma_map_page(dev, dest, 0, len, DMA_BIDIRECTIONAL);
dma_src[0] = dma_map_page(dev, src, 0, len, DMA_TO_DEVICE);
tx = dma->device_prep_dma_pq(chan, dma_dest, dma_src, 1, &coef,
async_tx_submit(chan, tx, submit);
return tx;
}
+
+ /* could not get a descriptor, unmap and fall through to
+ * the synchronous path
+ */
+ dma_unmap_page(dev, dma_dest[1], len, DMA_BIDIRECTIONAL);
+ dma_unmap_page(dev, dma_src[0], len, DMA_TO_DEVICE);
}
/* no channel available, or failed to allocate a descriptor, so
}
static struct dma_async_tx_descriptor *
-__2data_recov_4(size_t bytes, int faila, int failb, struct page **blocks,
- struct async_submit_ctl *submit)
+__2data_recov_4(int disks, size_t bytes, int faila, int failb,
+ struct page **blocks, struct async_submit_ctl *submit)
{
struct dma_async_tx_descriptor *tx = NULL;
struct page *p, *q, *a, *b;
void *cb_param = submit->cb_param;
void *scribble = submit->scribble;
- p = blocks[4-2];
- q = blocks[4-1];
+ p = blocks[disks-2];
+ q = blocks[disks-1];
a = blocks[faila];
b = blocks[failb];
srcs[1] = q;
coef[0] = raid6_gfexi[failb-faila];
coef[1] = raid6_gfinv[raid6_gfexp[faila]^raid6_gfexp[failb]];
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_sum_product(b, srcs, coef, bytes, submit);
/* Dy = P+Pxy+Dx */
}
static struct dma_async_tx_descriptor *
-__2data_recov_5(size_t bytes, int faila, int failb, struct page **blocks,
- struct async_submit_ctl *submit)
+__2data_recov_5(int disks, size_t bytes, int faila, int failb,
+ struct page **blocks, struct async_submit_ctl *submit)
{
struct dma_async_tx_descriptor *tx = NULL;
struct page *p, *q, *g, *dp, *dq;
dma_async_tx_callback cb_fn = submit->cb_fn;
void *cb_param = submit->cb_param;
void *scribble = submit->scribble;
- int uninitialized_var(good);
- int i;
+ int good_srcs, good, i;
- for (i = 0; i < 3; i++) {
+ good_srcs = 0;
+ good = -1;
+ for (i = 0; i < disks-2; i++) {
+ if (blocks[i] == NULL)
+ continue;
if (i == faila || i == failb)
continue;
- else {
- good = i;
- break;
- }
+ good = i;
+ good_srcs++;
}
- BUG_ON(i >= 3);
+ BUG_ON(good_srcs > 1);
- p = blocks[5-2];
- q = blocks[5-1];
+ p = blocks[disks-2];
+ q = blocks[disks-1];
g = blocks[good];
/* Compute syndrome with zero for the missing data pages
dp = blocks[faila];
dq = blocks[failb];
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_memcpy(dp, g, 0, 0, bytes, submit);
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_mult(dq, g, raid6_gfexp[good], bytes, submit);
/* compute P + Pxy */
srcs[0] = dp;
srcs[1] = p;
- init_async_submit(submit, ASYNC_TX_XOR_DROP_DST, tx, NULL, NULL,
- scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx,
+ NULL, NULL, scribble);
tx = async_xor(dp, srcs, 0, 2, bytes, submit);
/* compute Q + Qxy */
srcs[0] = dq;
srcs[1] = q;
- init_async_submit(submit, ASYNC_TX_XOR_DROP_DST, tx, NULL, NULL,
- scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx,
+ NULL, NULL, scribble);
tx = async_xor(dq, srcs, 0, 2, bytes, submit);
/* Dx = A*(P+Pxy) + B*(Q+Qxy) */
srcs[1] = dq;
coef[0] = raid6_gfexi[failb-faila];
coef[1] = raid6_gfinv[raid6_gfexp[faila]^raid6_gfexp[failb]];
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_sum_product(dq, srcs, coef, bytes, submit);
/* Dy = P+Pxy+Dx */
* delta p and delta q
*/
dp = blocks[faila];
- blocks[faila] = (void *)raid6_empty_zero_page;
+ blocks[faila] = NULL;
blocks[disks-2] = dp;
dq = blocks[failb];
- blocks[failb] = (void *)raid6_empty_zero_page;
+ blocks[failb] = NULL;
blocks[disks-1] = dq;
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_gen_syndrome(blocks, 0, disks, bytes, submit);
/* Restore pointer table */
/* compute P + Pxy */
srcs[0] = dp;
srcs[1] = p;
- init_async_submit(submit, ASYNC_TX_XOR_DROP_DST, tx, NULL, NULL,
- scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx,
+ NULL, NULL, scribble);
tx = async_xor(dp, srcs, 0, 2, bytes, submit);
/* compute Q + Qxy */
srcs[0] = dq;
srcs[1] = q;
- init_async_submit(submit, ASYNC_TX_XOR_DROP_DST, tx, NULL, NULL,
- scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx,
+ NULL, NULL, scribble);
tx = async_xor(dq, srcs, 0, 2, bytes, submit);
/* Dx = A*(P+Pxy) + B*(Q+Qxy) */
srcs[1] = dq;
coef[0] = raid6_gfexi[failb-faila];
coef[1] = raid6_gfinv[raid6_gfexp[faila]^raid6_gfexp[failb]];
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_sum_product(dq, srcs, coef, bytes, submit);
/* Dy = P+Pxy+Dx */
async_raid6_2data_recov(int disks, size_t bytes, int faila, int failb,
struct page **blocks, struct async_submit_ctl *submit)
{
+ int non_zero_srcs, i;
+
BUG_ON(faila == failb);
if (failb < faila)
swap(faila, failb);
*/
if (!submit->scribble) {
void **ptrs = (void **) blocks;
- int i;
async_tx_quiesce(&submit->depend_tx);
for (i = 0; i < disks; i++)
- ptrs[i] = page_address(blocks[i]);
+ if (blocks[i] == NULL)
+ ptrs[i] = (void *) raid6_empty_zero_page;
+ else
+ ptrs[i] = page_address(blocks[i]);
raid6_2data_recov(disks, bytes, faila, failb, ptrs);
return NULL;
}
- switch (disks) {
- case 4:
+ non_zero_srcs = 0;
+ for (i = 0; i < disks-2 && non_zero_srcs < 4; i++)
+ if (blocks[i])
+ non_zero_srcs++;
+ switch (non_zero_srcs) {
+ case 0:
+ case 1:
+ /* There must be at least 2 sources - the failed devices. */
+ BUG();
+
+ case 2:
/* dma devices do not uniformly understand a zero source pq
* operation (in contrast to the synchronous case), so
- * explicitly handle the 4 disk special case
+ * explicitly handle the special case of a 4 disk array with
+ * both data disks missing.
*/
- return __2data_recov_4(bytes, faila, failb, blocks, submit);
- case 5:
+ return __2data_recov_4(disks, bytes, faila, failb, blocks, submit);
+ case 3:
/* dma devices do not uniformly understand a single
* source pq operation (in contrast to the synchronous
- * case), so explicitly handle the 5 disk special case
+ * case), so explicitly handle the special case of a 5 disk
+ * array with 2 of 3 data disks missing.
*/
- return __2data_recov_5(bytes, faila, failb, blocks, submit);
+ return __2data_recov_5(disks, bytes, faila, failb, blocks, submit);
default:
return __2data_recov_n(disks, bytes, faila, failb, blocks, submit);
}
dma_async_tx_callback cb_fn = submit->cb_fn;
void *cb_param = submit->cb_param;
void *scribble = submit->scribble;
+ int good_srcs, good, i;
struct page *srcs[2];
pr_debug("%s: disks: %d len: %zu\n", __func__, disks, bytes);
*/
if (!scribble) {
void **ptrs = (void **) blocks;
- int i;
async_tx_quiesce(&submit->depend_tx);
for (i = 0; i < disks; i++)
- ptrs[i] = page_address(blocks[i]);
+ if (blocks[i] == NULL)
+ ptrs[i] = (void*)raid6_empty_zero_page;
+ else
+ ptrs[i] = page_address(blocks[i]);
raid6_datap_recov(disks, bytes, faila, ptrs);
return NULL;
}
+ good_srcs = 0;
+ good = -1;
+ for (i = 0; i < disks-2; i++) {
+ if (i == faila)
+ continue;
+ if (blocks[i]) {
+ good = i;
+ good_srcs++;
+ if (good_srcs > 1)
+ break;
+ }
+ }
+ BUG_ON(good_srcs == 0);
+
p = blocks[disks-2];
q = blocks[disks-1];
* Use the dead data page as temporary storage for delta q
*/
dq = blocks[faila];
- blocks[faila] = (void *)raid6_empty_zero_page;
+ blocks[faila] = NULL;
blocks[disks-1] = dq;
- /* in the 4 disk case we only need to perform a single source
- * multiplication
+ /* in the 4-disk case we only need to perform a single source
+ * multiplication with the one good data block.
*/
- if (disks == 4) {
- int good = faila == 0 ? 1 : 0;
+ if (good_srcs == 1) {
struct page *g = blocks[good];
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL,
+ scribble);
tx = async_memcpy(p, g, 0, 0, bytes, submit);
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL,
+ scribble);
tx = async_mult(dq, g, raid6_gfexp[good], bytes, submit);
} else {
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL,
+ scribble);
tx = async_gen_syndrome(blocks, 0, disks, bytes, submit);
}
srcs[0] = dq;
srcs[1] = q;
- init_async_submit(submit, ASYNC_TX_XOR_DROP_DST, tx, NULL, NULL,
- scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx,
+ NULL, NULL, scribble);
tx = async_xor(dq, srcs, 0, 2, bytes, submit);
- init_async_submit(submit, 0, tx, NULL, NULL, scribble);
+ init_async_submit(submit, ASYNC_TX_FENCE, tx, NULL, NULL, scribble);
tx = async_mult(dq, dq, coef, bytes, submit);
srcs[0] = p;