qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Alexander Graf <agraf@suse.de>
To: Alexey Kardashevskiy <aik@ozlabs.ru>, qemu-devel@nongnu.org
Cc: Alex Williamson <alex.williamson@redhat.com>, qemu-ppc@nongnu.org
Subject: Re: [Qemu-devel] [RFC PATCH 06/10] spapr_rtas: Add Dynamic DMA windows (DDW) RTAS calls support
Date: Mon, 11 Aug 2014 13:51:39 +0200	[thread overview]
Message-ID: <53E8AE4B.9030500@suse.de> (raw)
In-Reply-To: <1406799254-25223-7-git-send-email-aik@ozlabs.ru>


On 31.07.14 11:34, Alexey Kardashevskiy wrote:
> This adds support for Dynamic DMA Windows (DDW) option defined by
> the SPAPR specification which allows to have additional DMA window(s)
> which can support page sizes other than 4K.
>
> The existing implementation of DDW in the guest tries to create one huge
> DMA window with 64K or 16MB pages and map the entire guest RAM to. If it
> succeeds, the guest switches to dma_direct_ops and never calls
> TCE hypercalls (H_PUT_TCE,...) again. This enables VFIO devices to use
> the entire RAM and not waste time on map/unmap.
>
> This adds 4 RTAS handlers:
> * ibm,query-pe-dma-window
> * ibm,create-pe-dma-window
> * ibm,remove-pe-dma-window
> * ibm,reset-pe-dma-window
> These are registered from type_init() callback.
>
> These RTAS handlers are implemented in a separate file to avoid polluting
> spapr_iommu.c with PHB.
>
> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
> ---
>   hw/ppc/Makefile.objs        |   3 +
>   hw/ppc/spapr_rtas_ddw.c     | 296 ++++++++++++++++++++++++++++++++++++++++++++
>   include/hw/pci-host/spapr.h |  18 +++
>   include/hw/ppc/spapr.h      |   6 +-
>   trace-events                |   4 +
>   5 files changed, 326 insertions(+), 1 deletion(-)
>   create mode 100644 hw/ppc/spapr_rtas_ddw.c
>
> diff --git a/hw/ppc/Makefile.objs b/hw/ppc/Makefile.objs
> index edd44d0..9773294 100644
> --- a/hw/ppc/Makefile.objs
> +++ b/hw/ppc/Makefile.objs
> @@ -7,6 +7,9 @@ obj-$(CONFIG_PSERIES) += spapr_pci.o
>   ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)
>   obj-y += spapr_pci_vfio.o
>   endif
> +ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES), yy)
> +obj-y += spapr_rtas_ddw.o
> +endif
>   # PowerPC 4xx boards
>   obj-y += ppc405_boards.o ppc4xx_devs.o ppc405_uc.o ppc440_bamboo.o
>   obj-y += ppc4xx_pci.o
> diff --git a/hw/ppc/spapr_rtas_ddw.c b/hw/ppc/spapr_rtas_ddw.c
> new file mode 100644
> index 0000000..943af2c
> --- /dev/null
> +++ b/hw/ppc/spapr_rtas_ddw.c
> @@ -0,0 +1,296 @@
> +/*
> + * QEMU sPAPR Dynamic DMA windows support
> + *
> + * Copyright (c) 2014 Alexey Kardashevskiy, IBM Corporation.
> + *
> + *  This program is free software; you can redistribute it and/or modify
> + *  it under the terms of the GNU General Public License as published by
> + *  the Free Software Foundation; either version 2 of the License,
> + *  or (at your option) any later version.
> + *
> + *  This program is distributed in the hope that it will be useful,
> + *  but WITHOUT ANY WARRANTY; without even the implied warranty of
> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + *  GNU General Public License for more details.
> + *
> + *  You should have received a copy of the GNU General Public License
> + *  along with this program; if not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#include "hw/ppc/spapr.h"
> +#include "hw/pci-host/spapr.h"
> +#include "trace.h"
> +
> +static inline uint32_t spapr_iommu_fixmask(uint32_t cur_mask,
> +                                           struct ppc_one_seg_page_size *sps,
> +                                           uint32_t query_mask,
> +                                           int shift,
> +                                           uint32_t add_mask)
> +{
> +    if ((sps->page_shift == shift) && (query_mask & add_mask)) {
> +        cur_mask |= add_mask;
> +    }
> +    return cur_mask;
> +}
> +
> +static void rtas_ibm_query_pe_dma_window(PowerPCCPU *cpu,
> +                                         sPAPREnvironment *spapr,
> +                                         uint32_t token, uint32_t nargs,
> +                                         target_ulong args,
> +                                         uint32_t nret, target_ulong rets)
> +{
> +    CPUPPCState *env = &cpu->env;
> +    sPAPRPHBState *sphb;
> +    sPAPRPHBClass *spc;
> +    uint64_t buid;
> +    uint32_t addr, pgmask = 0;
> +    uint32_t windows_available = 0, page_size_mask = 0;
> +    long ret, i;
> +
> +    if ((nargs != 3) || (nret != 5)) {
> +        goto param_error_exit;
> +    }
> +
> +    buid = ((uint64_t)rtas_ld(args, 1) << 32) | rtas_ld(args, 2);
> +    addr = rtas_ld(args, 0);
> +    sphb = spapr_pci_find_phb(spapr, buid);
> +    if (!sphb) {
> +        goto param_error_exit;
> +    }
> +
> +    spc = SPAPR_PCI_HOST_BRIDGE_GET_CLASS(sphb);
> +    if (!spc->ddw_query) {
> +        goto hw_error_exit;
> +    }
> +
> +    ret = spc->ddw_query(sphb, &windows_available, &page_size_mask);
> +    trace_spapr_iommu_ddw_query(buid, addr, windows_available,
> +                                page_size_mask, pgmask, ret);
> +    if (ret) {
> +        goto hw_error_exit;
> +    }
> +
> +    /* DBG! */
> +    if (!(page_size_mask & DDW_PGSIZE_16M)) {
> +        goto hw_error_exit;
> +    }
> +
> +    /* Work out biggest possible page size */
> +    for (i = 0; i < PPC_PAGE_SIZES_MAX_SZ; i++) {
> +        int j;
> +        struct ppc_one_seg_page_size *sps = &env->sps.sps[i];
> +        const struct { int shift; uint32_t mask; } masks[] = {
> +            { 12, DDW_PGSIZE_4K },
> +            { 16, DDW_PGSIZE_64K },
> +            { 24, DDW_PGSIZE_16M },
> +            { 25, DDW_PGSIZE_32M },
> +            { 26, DDW_PGSIZE_64M },
> +            { 27, DDW_PGSIZE_128M },
> +            { 28, DDW_PGSIZE_256M },
> +            { 34, DDW_PGSIZE_16G },
> +        };
> +        for (j = 0; j < ARRAY_SIZE(masks); ++j) {
> +            pgmask = spapr_iommu_fixmask(pgmask, sps, page_size_mask,
> +                                         masks[j].shift, masks[j].mask);
> +        }
> +    }
> +
> +    rtas_st(rets, 0, RTAS_OUT_SUCCESS);
> +    rtas_st(rets, 1, windows_available);
> +    /* Return maximum number as all RAM was 4K pages */
> +    rtas_st(rets, 2, ram_size >> SPAPR_TCE_PAGE_SHIFT);
> +    rtas_st(rets, 3, pgmask);
> +    rtas_st(rets, 4, pgmask); /* DMA migration mask */
> +    return;
> +
> +hw_error_exit:
> +    rtas_st(rets, 0, RTAS_OUT_HW_ERROR);
> +    return;
> +
> +param_error_exit:
> +    rtas_st(rets, 0, RTAS_OUT_PARAM_ERROR);
> +}
> +
> +static void rtas_ibm_create_pe_dma_window(PowerPCCPU *cpu,
> +                                          sPAPREnvironment *spapr,
> +                                          uint32_t token, uint32_t nargs,
> +                                          target_ulong args,
> +                                          uint32_t nret, target_ulong rets)
> +{
> +    sPAPRPHBState *sphb;
> +    sPAPRPHBClass *spc;
> +    sPAPRTCETable *tcet = NULL;
> +    uint32_t addr, page_shift, window_shift, liobn;
> +    uint64_t buid;
> +    long ret;
> +
> +    if ((nargs != 5) || (nret != 4)) {
> +        goto param_error_exit;
> +    }
> +
> +    buid = ((uint64_t)rtas_ld(args, 1) << 32) | rtas_ld(args, 2);
> +    addr = rtas_ld(args, 0);
> +    sphb = spapr_pci_find_phb(spapr, buid);
> +    if (!sphb) {
> +        goto param_error_exit;
> +    }
> +
> +    spc = SPAPR_PCI_HOST_BRIDGE_GET_CLASS(sphb);
> +    if (!spc->ddw_create) {
> +        goto hw_error_exit;
> +    }
> +
> +    page_shift = rtas_ld(args, 3);
> +    window_shift = rtas_ld(args, 4);
> +    liobn = sphb->dma_liobn + 0x10000;

What offset is this?

> +
> +    ret = spc->ddw_create(sphb, page_shift, window_shift, liobn, &tcet);
> +    trace_spapr_iommu_ddw_create(buid, addr, 1 << page_shift,
> +                                 1 << window_shift,

1ULL? Otherwise 16G pages (and windows) won't work.


Alex

  reply	other threads:[~2014-08-11 11:51 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-31  9:34 [Qemu-devel] [RFC PATCH 00/10] spapr: vfio: Enable Dynamic DMA windows (DDW) Alexey Kardashevskiy
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 01/10] qom: Make object_child_foreach safe for objects removal Alexey Kardashevskiy
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 02/10] spapr_iommu: Disable in-kernel IOMMU tables for >4GB windows Alexey Kardashevskiy
2014-08-12  1:17   ` David Gibson
2014-08-12  7:32     ` Alexey Kardashevskiy
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 03/10] spapr_pci: Make find_phb()/find_dev() public Alexey Kardashevskiy
2014-08-11 11:39   ` Alexander Graf
2014-08-11 14:56     ` Alexey Kardashevskiy
2014-08-11 17:16       ` Alexander Graf
2014-08-12  1:19   ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 04/10] spapr_iommu: Make spapr_tce_find_by_liobn() public Alexey Kardashevskiy
2014-08-12  1:19   ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 05/10] linux headers update for DDW Alexey Kardashevskiy
2014-08-12  1:20   ` David Gibson
2014-08-12  7:16     ` Alexey Kardashevskiy
2014-08-13  3:23       ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 06/10] spapr_rtas: Add Dynamic DMA windows (DDW) RTAS calls support Alexey Kardashevskiy
2014-08-11 11:51   ` Alexander Graf [this message]
2014-08-11 15:34     ` Alexey Kardashevskiy
2014-08-12  1:45   ` David Gibson
2014-08-12  7:25     ` Alexey Kardashevskiy
2014-08-13  3:27       ` David Gibson
2014-08-14  8:29         ` Alexey Kardashevskiy
2014-08-15  0:04           ` David Gibson
2014-08-15  3:09             ` Alexey Kardashevskiy
2014-08-15  4:20               ` David Gibson
2014-08-15  5:27                 ` Alexey Kardashevskiy
2014-08-15  5:30                   ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 07/10] spapr: Add "ddw" machine option Alexey Kardashevskiy
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 08/10] spapr_pci: Enable DDW Alexey Kardashevskiy
2014-08-11 11:59   ` Alexander Graf
2014-08-11 15:26     ` Alexey Kardashevskiy
2014-08-11 17:29       ` Alexander Graf
2014-08-12  0:13         ` Alexey Kardashevskiy
2014-08-12  3:59           ` Alexey Kardashevskiy
2014-08-12  9:36             ` Alexander Graf
2014-08-12  2:10   ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 09/10] spapr_pci_vfio: " Alexey Kardashevskiy
2014-08-11 12:02   ` Alexander Graf
2014-08-11 15:01     ` Alexey Kardashevskiy
2014-08-11 17:30       ` Alexander Graf
2014-08-12  0:03         ` Alexey Kardashevskiy
2014-08-12  9:37           ` Alexander Graf
2014-08-12 15:10             ` Alexey Kardashevskiy
2014-08-12 15:28               ` Alexander Graf
2014-08-13  0:18                 ` Alexey Kardashevskiy
2014-08-14 13:38                   ` Alexander Graf
2014-08-15  0:09                     ` David Gibson
2014-08-15  3:22                       ` Alexey Kardashevskiy
2014-08-15  3:16                     ` Alexey Kardashevskiy
2014-08-15  7:37                       ` Alexander Graf
2014-08-12  2:14   ` David Gibson
2014-07-31  9:34 ` [Qemu-devel] [RFC PATCH 10/10] vfio: Enable DDW ioctls to VFIO IOMMU driver Alexey Kardashevskiy
2014-08-05  1:30 ` [Qemu-devel] [RFC PATCH 00/10] spapr: vfio: Enable Dynamic DMA windows (DDW) Alexey Kardashevskiy
2014-08-10 23:50   ` Alexey Kardashevskiy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=53E8AE4B.9030500@suse.de \
    --to=agraf@suse.de \
    --cc=aik@ozlabs.ru \
    --cc=alex.williamson@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).