[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 3/5] ide: move buffered DMA cancel to core
From: |
Paolo Bonzini |
Subject: |
Re: [Qemu-devel] [PATCH 3/5] ide: move buffered DMA cancel to core |
Date: |
Tue, 19 Jan 2016 12:50:43 +0100 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.5.0 |
On 19/01/2016 05:51, John Snow wrote:
> Buffered DMA cancellation was added to ATAPI devices and implemented
> for the BMDMA HBA. Move the code over to common IDE code and allow
> it to be used for any HBA.
>
> Signed-off-by: John Snow <address@hidden>
> ---
> hw/ide/core.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
> hw/ide/internal.h | 1 +
> hw/ide/pci.c | 36 +-----------------------------------
> 3 files changed, 47 insertions(+), 35 deletions(-)
>
> diff --git a/hw/ide/core.c b/hw/ide/core.c
> index 75486c2..5d81840 100644
> --- a/hw/ide/core.c
> +++ b/hw/ide/core.c
> @@ -608,6 +608,51 @@ BlockAIOCB *ide_buffered_readv(IDEState *s, int64_t
> sector_num,
> return aioreq;
> }
>
> +/**
> + * Cancel all pending DMA requests.
> + * Any buffered DMA requests are instantly canceled,
> + * but any pending unbuffered DMA requests must be waited on.
> + */
> +void ide_cancel_dma_sync(IDEState *s)
> +{
> + IDEBufferedRequest *req;
> +
> + /* First invoke the callbacks of all buffered requests
> + * and flag those requests as orphaned. Ideally there
> + * are no unbuffered (Scatter Gather DMA Requests or
> + * write requests) pending and we can avoid to drain. */
> + QLIST_FOREACH(req, &s->buffered_requests, list) {
> + if (!req->orphaned) {
> +#ifdef DEBUG_IDE
> + printf("%s: invoking cb %p of buffered request %p with"
> + " -ECANCELED\n", __func__, req->original_cb, req);
> +#endif
> + req->original_cb(req->original_opaque, -ECANCELED);
> + }
> + req->orphaned = true;
> + }
> +
> + /*
> + * We can't cancel Scatter Gather DMA in the middle of the
> + * operation or a partial (not full) DMA transfer would reach
> + * the storage so we wait for completion instead (we beahve
> + * like if the DMA was completed by the time the guest trying
> + * to cancel dma with bmdma_cmd_writeb with BM_CMD_START not
> + * set).
> + *
> + * In the future we'll be able to safely cancel the I/O if the
> + * whole DMA operation will be submitted to disk with a single
> + * aio operation with preadv/pwritev.
> + */
> + if (s->bus->dma->aiocb) {
> +#ifdef DEBUG_IDE
> + printf("%s: draining all remaining requests", __func__);
> +#endif
> + blk_drain_all();
As a separate patch you can change this to blk_drain(s->blk), which is
already an improvement.
Paolo
> + assert(s->bus->dma->aiocb == NULL);
> + }
> +}
> +
> static void ide_sector_read(IDEState *s);
>
> static void ide_sector_read_cb(void *opaque, int ret)
> diff --git a/hw/ide/internal.h b/hw/ide/internal.h
> index 2d1e2d2..86bde26 100644
> --- a/hw/ide/internal.h
> +++ b/hw/ide/internal.h
> @@ -586,6 +586,7 @@ BlockAIOCB *ide_issue_trim(BlockBackend *blk,
> BlockAIOCB *ide_buffered_readv(IDEState *s, int64_t sector_num,
> QEMUIOVector *iov, int nb_sectors,
> BlockCompletionFunc *cb, void *opaque);
> +void ide_cancel_dma_sync(IDEState *s);
>
> /* hw/ide/atapi.c */
> void ide_atapi_cmd(IDEState *s);
> diff --git a/hw/ide/pci.c b/hw/ide/pci.c
> index 37dbc29..6b780b8 100644
> --- a/hw/ide/pci.c
> +++ b/hw/ide/pci.c
> @@ -233,41 +233,7 @@ void bmdma_cmd_writeb(BMDMAState *bm, uint32_t val)
> /* Ignore writes to SSBM if it keeps the old value */
> if ((val & BM_CMD_START) != (bm->cmd & BM_CMD_START)) {
> if (!(val & BM_CMD_START)) {
> - /* First invoke the callbacks of all buffered requests
> - * and flag those requests as orphaned. Ideally there
> - * are no unbuffered (Scatter Gather DMA Requests or
> - * write requests) pending and we can avoid to drain. */
> - IDEBufferedRequest *req;
> - IDEState *s = idebus_active_if(bm->bus);
> - QLIST_FOREACH(req, &s->buffered_requests, list) {
> - if (!req->orphaned) {
> -#ifdef DEBUG_IDE
> - printf("%s: invoking cb %p of buffered request %p with"
> - " -ECANCELED\n", __func__, req->original_cb, req);
> -#endif
> - req->original_cb(req->original_opaque, -ECANCELED);
> - }
> - req->orphaned = true;
> - }
> - /*
> - * We can't cancel Scatter Gather DMA in the middle of the
> - * operation or a partial (not full) DMA transfer would reach
> - * the storage so we wait for completion instead (we beahve
> - * like if the DMA was completed by the time the guest trying
> - * to cancel dma with bmdma_cmd_writeb with BM_CMD_START not
> - * set).
> - *
> - * In the future we'll be able to safely cancel the I/O if the
> - * whole DMA operation will be submitted to disk with a single
> - * aio operation with preadv/pwritev.
> - */
> - if (bm->bus->dma->aiocb) {
> -#ifdef DEBUG_IDE
> - printf("%s: draining all remaining requests", __func__);
> -#endif
> - blk_drain_all();
> - assert(bm->bus->dma->aiocb == NULL);
> - }
> + ide_cancel_dma_sync(idebus_active_if(bm->bus));
> bm->status &= ~BM_STATUS_DMAING;
> } else {
> bm->cur_addr = bm->addr;
>