[PATCH] dma driver: fix potential oom issue of fsldma

From: Xuelin Shi <xuelin@freescale.com> missing unmap sources and destinations while doing dequeue. Signed-off-by: Xuelin Shi <xuelin@nxp.com> --- drivers/dma/fsldma.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c i

RE: [PATCH v5] dmaengine: Driver support for FSL RaidEngine device.

Hi, Please see my comments inline. Thanks, Shi xuelin -Original Message- From: Vinod Koul [mailto:vinod.k...@intel.com] Sent: Wednesday, February 11, 2015 09:06 To: Shi Xuelin-B29237 Cc: dan.j.willi...@intel.com; dmaeng...@vger.kernel.org; linuxppc- d...@lists.ozlabs.org; Rai

RE: [PATCH v3] dmaengine: driver support for FSL RaidEngine device.

Hi Dan Vinod, I have sent out the v4 of this patch and not received any further feedback yet. This patch looks ruled out from the patchwork. https://patchwork.kernel.org/project/linux-dmaengine/list/?page=2 So do you know what happened to this patch? Thanks, Xuelin Shi -Original

RE: [PATCH v3] dmaengine: driver support for FSL RaidEngine device.

Yes, depend on !ASYNC_TX_CHANNEL_SWITCH is better since fsldma selects this condition. Thanks, Xuelin Shi -Original Message- From: Dan Williams [mailto:dan.j.willi...@intel.com] Sent: 2014年4月15日 8:30 To: Shi Xuelin-B29237 Cc: Koul, Vinod; andriy.shevche...@intel.com; dmaeng

RE: [PATCH v3] dmaengine: driver support for FSL RaidEngine device.

Hi Dan, fsl dma device and fsl raid device are two differenct devices that both provide async_memcpy capability, so I use !FSL_DMA to disable the fsl dma device. That's to say, either select fsldma device, either fsl raid device. Thanks, Xuelin Shi -Original Message- From

RE: [PATCH v2] fix wrong usage of dmaengine_unmap_put in async_xxx

Hi Dan, With https://patchwork.kernel.org/patch/3863711/ applied, the issue disappeared. Thanks, Xuelin Shi -Original Message- From: dan.j.willi...@gmail.com [mailto:dan.j.willi...@gmail.com] On Behalf Of Dan Williams Sent: 2014年4月10日 12:02 To: Shi Xuelin-B29237 Cc: Koul, Vinod; dmaeng

RE: [PATCH v2] fix wrong usage of dmaengine_unmap_put in async_xxx

Hi Dan vinod, Have you got a chance to see this patch? Also what's your comment about the RaidEngine driver patch? Thanks, Xuelin Shi -Original Message- From: xuelin@freescale.com [mailto:xuelin@freescale.com] Sent: 2014年3月20日 16:16 To: vinod.k...@intel.com; dan.j.willi

RE: [PATCH] fix dmaengine_unmap failure.

Hi Dan, I'm OK to save memory here. I'd like to send v2. Thanks, Xuelin Shi -Original Message- From: Dan Williams [mailto:dan.j.willi...@intel.com] Sent: 2014年3月19日 23:44 To: Shi Xuelin-B29237 Cc: Vinod Koul; linuxppc-dev; dmaeng...@vger.kernel.org Subject: Re: [PATCH] fix

RE: [PATCH v2] fix wrong usage of dmaengine_unmap_put in async_xxx

Subject: Re: [PATCH v2] fix wrong usage of dmaengine_unmap_put in async_xxx On Thu, 2014-03-20 at 16:16 +0800, xuelin@freescale.com wrote: From: Xuelin Shi xuelin@freescale.com dmaengine_unmap_put does below two things: a) unmap pages for srcs and dests b) free unmap struct

RE: [PATCH] fix dmaengine_unmap failure.

with the unmap structure instead of computing it again. By this way, it is guaranteed that the unmap is the same when we get and put the unmap data. BTW: the mempool is just used to manage the struct unmap, not the pages. Thanks, Xuelin Shi -Original Message- From: Dan Williams