diff options
author | Dan Williams <dan.j.williams@intel.com> | 2013-11-13 10:37:36 -0800 |
---|---|---|
committer | Dan Williams <dan.j.williams@intel.com> | 2013-11-14 11:04:41 -0800 |
commit | 21e96c7313486390c694919522a76dfea0a86c59 (patch) | |
tree | 211439af0431a256def810a915e37ee6519ae4a6 /drivers/dma/ioat/dma_v3.c | |
parent | 5d48b9b5d80e3aa38a5161565398b1e48a650573 (diff) | |
download | op-kernel-dev-21e96c7313486390c694919522a76dfea0a86c59.zip op-kernel-dev-21e96c7313486390c694919522a76dfea0a86c59.tar.gz |
ioatdma: fix selection of 16 vs 8 source path
When performing continuations there are implied sources that need to be
added to the source count. Quoting dma_set_maxpq:
/* dma_maxpq - reduce maxpq in the face of continued operations
* @dma - dma device with PQ capability
* @flags - to check if DMA_PREP_CONTINUE and DMA_PREP_PQ_DISABLE_P are set
*
* When an engine does not support native continuation we need 3 extra
* source slots to reuse P and Q with the following coefficients:
* 1/ {00} * P : remove P from Q', but use it as a source for P'
* 2/ {01} * Q : use Q to continue Q' calculation
* 3/ {00} * Q : subtract Q from P' to cancel (2)
*
* In the case where P is disabled we only need 1 extra source:
* 1/ {01} * Q : use Q to continue Q' calculation
*/
...fix the selection of the 16 source path to take these implied sources
into account.
Note this also kills the BUG_ON(src_cnt < 9) check in
__ioat3_prep_pq16_lock(). Besides not accounting for implied sources
the check is redundant given we already made the path selection.
Cc: <stable@vger.kernel.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Diffstat (limited to 'drivers/dma/ioat/dma_v3.c')
-rw-r--r-- | drivers/dma/ioat/dma_v3.c | 30 |
1 files changed, 15 insertions, 15 deletions
diff --git a/drivers/dma/ioat/dma_v3.c b/drivers/dma/ioat/dma_v3.c index f26a35d..0284908 100644 --- a/drivers/dma/ioat/dma_v3.c +++ b/drivers/dma/ioat/dma_v3.c @@ -951,9 +951,6 @@ __ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result, u8 op; int i, s, idx, num_descs; - /* this function only handles src_cnt 9 - 16 */ - BUG_ON(src_cnt < 9); - /* this function is only called with 9-16 sources */ op = result ? IOAT_OP_PQ_VAL_16S : IOAT_OP_PQ_16S; @@ -1039,13 +1036,21 @@ __ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result, return &desc->txd; } +static int src_cnt_flags(unsigned int src_cnt, unsigned long flags) +{ + if (dmaf_p_disabled_continue(flags)) + return src_cnt + 1; + else if (dmaf_continue(flags)) + return src_cnt + 3; + else + return src_cnt; +} + static struct dma_async_tx_descriptor * ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, unsigned int src_cnt, const unsigned char *scf, size_t len, unsigned long flags) { - struct dma_device *dma = chan->device; - /* specify valid address for disabled result */ if (flags & DMA_PREP_PQ_DISABLE_P) dst[0] = dst[1]; @@ -1065,7 +1070,7 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, single_source_coef[0] = scf[0]; single_source_coef[1] = 0; - return (src_cnt > 8) && (dma->max_pq > 8) ? + return src_cnt_flags(src_cnt, flags) > 8 ? __ioat3_prep_pq16_lock(chan, NULL, dst, single_source, 2, single_source_coef, len, flags) : @@ -1073,7 +1078,7 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, single_source_coef, len, flags); } else { - return (src_cnt > 8) && (dma->max_pq > 8) ? + return src_cnt_flags(src_cnt, flags) > 8 ? __ioat3_prep_pq16_lock(chan, NULL, dst, src, src_cnt, scf, len, flags) : __ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt, @@ -1086,8 +1091,6 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, unsigned int src_cnt, const unsigned char *scf, size_t len, enum sum_check_flags *pqres, unsigned long flags) { - struct dma_device *dma = chan->device; - /* specify valid address for disabled result */ if (flags & DMA_PREP_PQ_DISABLE_P) pq[0] = pq[1]; @@ -1099,7 +1102,7 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, */ *pqres = 0; - return (src_cnt > 8) && (dma->max_pq > 8) ? + return src_cnt_flags(src_cnt, flags) > 8 ? __ioat3_prep_pq16_lock(chan, pqres, pq, src, src_cnt, scf, len, flags) : __ioat3_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len, @@ -1110,7 +1113,6 @@ static struct dma_async_tx_descriptor * ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, unsigned int src_cnt, size_t len, unsigned long flags) { - struct dma_device *dma = chan->device; unsigned char scf[src_cnt]; dma_addr_t pq[2]; @@ -1119,7 +1121,7 @@ ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, flags |= DMA_PREP_PQ_DISABLE_Q; pq[1] = dst; /* specify valid address for disabled result */ - return (src_cnt > 8) && (dma->max_pq > 8) ? + return src_cnt_flags(src_cnt, flags) > 8 ? __ioat3_prep_pq16_lock(chan, NULL, pq, src, src_cnt, scf, len, flags) : __ioat3_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len, @@ -1131,7 +1133,6 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, unsigned int src_cnt, size_t len, enum sum_check_flags *result, unsigned long flags) { - struct dma_device *dma = chan->device; unsigned char scf[src_cnt]; dma_addr_t pq[2]; @@ -1145,8 +1146,7 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, flags |= DMA_PREP_PQ_DISABLE_Q; pq[1] = pq[0]; /* specify valid address for disabled result */ - - return (src_cnt > 8) && (dma->max_pq > 8) ? + return src_cnt_flags(src_cnt, flags) > 8 ? __ioat3_prep_pq16_lock(chan, result, pq, &src[1], src_cnt - 1, scf, len, flags) : __ioat3_prep_pq_lock(chan, result, pq, &src[1], src_cnt - 1, |