From: Bharat Bhushan Subject: [PATCH] Using alloc_coherent for caam job rings Date: Wed, 27 Jun 2012 10:58:32 +0530 Message-ID: <1340774912-19042-1-git-send-email-bharat.bhushan@freescale.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Bharat Bhushan , Bharat Bhushan To: , , Return-path: Received: from tx2ehsobe001.messaging.microsoft.com ([65.55.88.11]:8446 "EHLO tx2outboundpool.messaging.microsoft.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750708Ab2F0FX7 (ORCPT ); Wed, 27 Jun 2012 01:23:59 -0400 Sender: linux-crypto-owner@vger.kernel.org List-ID: From: Bharat Bhushan The caam job rings (input/output job ring) are allocated using dma_map_single(). These job rings can be visualized as the ring buffers in which the jobs are en-queued/de-queued. The s/w enqueues the jobs in input job ring which h/w dequeues and after processing it copies the jobs in output job ring. Software then de-queues the job from output ring. Using dma_map/unmap_single() is not preferred way to allocate memory for this type of requirements because this adds un-necessary complexity. Example, if bounce buffer (SWIOTLB) will get used then to make any change visible in this memory to other processing unit requires dmap_unmap_single() or dma_sync_single_for_cpu/device(). The dma_unmap_single() can not be used as this will free the bounce buffer, this will require changing the job rings on running system and I seriously doubt that it will be not possible or very complex to implement. Also using dma_sync_single_for_cpu/device() will also add unnecessary complexity. The simple and preferred way is using dma_alloc_coherent() for these type of memory requirements. This resolves the Linux boot crash issue when "swiotlb=force" is set in bootargs on systems which have memory more than 4G. Signed-off-by: Bharat Bhushan --- This patch is based on "git://git.freescale.com/crypto/cryptodev.git master" which have Kim Phillip's "[PATCH 00/20] crypto: caam - driver updates" patches, which are already sent uptream. drivers/crypto/caam/jr.c | 45 +++++++++------------------------------------ 1 files changed, 9 insertions(+), 36 deletions(-) diff --git a/drivers/crypto/caam/jr.c b/drivers/crypto/caam/jr.c index 7074a1a..53c8c51 100644 --- a/drivers/crypto/caam/jr.c +++ b/drivers/crypto/caam/jr.c @@ -339,10 +339,11 @@ static int caam_jr_init(struct device *dev) if (error) return error; - jrp->inpring = kzalloc(sizeof(dma_addr_t) * JOBR_DEPTH, - GFP_KERNEL | GFP_DMA); - jrp->outring = kzalloc(sizeof(struct jr_outentry) * - JOBR_DEPTH, GFP_KERNEL | GFP_DMA); + jrp->inpring = dma_alloc_coherent(dev, sizeof(dma_addr_t) * JOBR_DEPTH, + &inpbusaddr, GFP_KERNEL); + + jrp->outring = dma_alloc_coherent(dev, sizeof(struct jr_outentry) * + JOBR_DEPTH, &outbusaddr, GFP_KERNEL); jrp->entinfo = kzalloc(sizeof(struct caam_jrentry_info) * JOBR_DEPTH, GFP_KERNEL); @@ -358,31 +359,6 @@ static int caam_jr_init(struct device *dev) jrp->entinfo[i].desc_addr_dma = !0; /* Setup rings */ - inpbusaddr = dma_map_single(dev, jrp->inpring, - sizeof(dma_addr_t) * JOBR_DEPTH, - DMA_BIDIRECTIONAL); - if (dma_mapping_error(dev, inpbusaddr)) { - dev_err(dev, "caam_jr_init(): can't map input ring\n"); - kfree(jrp->inpring); - kfree(jrp->outring); - kfree(jrp->entinfo); - return -EIO; - } - - outbusaddr = dma_map_single(dev, jrp->outring, - sizeof(struct jr_outentry) * JOBR_DEPTH, - DMA_BIDIRECTIONAL); - if (dma_mapping_error(dev, outbusaddr)) { - dev_err(dev, "caam_jr_init(): can't map output ring\n"); - dma_unmap_single(dev, inpbusaddr, - sizeof(dma_addr_t) * JOBR_DEPTH, - DMA_BIDIRECTIONAL); - kfree(jrp->inpring); - kfree(jrp->outring); - kfree(jrp->entinfo); - return -EIO; - } - jrp->inp_ring_write_index = 0; jrp->out_ring_read_index = 0; jrp->head = 0; @@ -426,13 +402,10 @@ int caam_jr_shutdown(struct device *dev) /* Free rings */ inpbusaddr = rd_reg64(&jrp->rregs->inpring_base); outbusaddr = rd_reg64(&jrp->rregs->outring_base); - dma_unmap_single(dev, outbusaddr, - sizeof(struct jr_outentry) * JOBR_DEPTH, - DMA_BIDIRECTIONAL); - dma_unmap_single(dev, inpbusaddr, sizeof(dma_addr_t) * JOBR_DEPTH, - DMA_BIDIRECTIONAL); - kfree(jrp->outring); - kfree(jrp->inpring); + dma_free_coherent(dev, sizeof(dma_addr_t) * JOBR_DEPTH, + jrp->inpring, inpbusaddr); + dma_free_coherent(dev, sizeof(struct jr_outentry) * JOBR_DEPTH, + jrp->outring, outbusaddr); kfree(jrp->entinfo); return ret; -- 1.7.0.4