| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CFD3932D.8F2A2%kadams@nvidia.com>
Date: Fri, 27 Jun 2014 18:41:58 -0700
From: Ken Adams <KAdams@...dia.com>
To: Ben Skeggs <skeggsb@...il.com>
CC: Alex Courbot <acourbot@...dia.com>,
Ben Skeggs <bskeggs@...hat.com>,
"linux-tegra@...r.kernel.org" <linux-tegra@...r.kernel.org>,
"nouveau@...ts.freedesktop.org" <nouveau@...ts.freedesktop.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>,
"gnurou@...il.com" <gnurou@...il.com>
Subject: Re: [PATCH v2] drm/gk20a: add BAR instance
On 6/27/14 8:56 PM, "Ben Skeggs" <skeggsb@...il.com> wrote:
>On Sat, Jun 28, 2014 at 4:51 AM, Ken Adams <KAdams@...dia.com> wrote:
>> quick note re: tegra and gpu bars...
>>
>> to this point we've explicitly avoided providing user-mode mappings due
>>to
>> power management issues, etc.
>> looks to me like this would allow such mappings. is that the case? are
>> there any paths which would require such mappings to function properly?
>What power management issues are you worried about in particular? We
>have these concerns on discrete cards too, when doing things like
>changing vram frequencies. TTM is able to kick out all userspace
>mappings, and clients will then block in the fault handler until it's
>safe - if they touch the mappings.
>
>Ben.
hi ben,
primarily it's the access problem you mentioned. managing those mappings,
and kicking them out at best adds to the latency to take down power/detach
busii and the like.
and, generally, there are very few (if any) cases where there isn't a
better way to manipulate the pixels than with the cpu :) i understand
there are plenty of paths i don't know about hereŠ and so i asked.
it's a solvable problem, of course. but especially in the mobile world it
can pop up unexpectedly. typically on someone's perf/power/stress tests :)
---
ken
>
>>
>> thanks
>> ---
>> ken
>>
>> p.s.: hello :)
>>
>> On 6/27/14 7:36 AM, "Alex Courbot" <acourbot@...dia.com> wrote:
>>
>>>GK20A's BAR is functionally identical to NVC0's, but do not support
>>>being ioremapped write-combined. Create a BAR instance for GK20A that
>>>reflect that state.
>>>
>>>Signed-off-by: Alexandre Courbot <acourbot@...dia.com>
>>>---
>>>Changes since v1:
>>>- Fix compilation warning due to missing cast
>>>
>>>Patch 1 of the series was ok and thus has not been resent.
>>>
>>> drivers/gpu/drm/nouveau/Makefile | 1 +
>>> drivers/gpu/drm/nouveau/core/engine/device/nve0.c | 2 +-
>>> drivers/gpu/drm/nouveau/core/include/subdev/bar.h | 1 +
>>> drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c | 54
>>>+++++++++++++++++++++++
>>> drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c | 6 +--
>>> drivers/gpu/drm/nouveau/core/subdev/bar/priv.h | 6 +++
>>> 6 files changed, 66 insertions(+), 4 deletions(-)
>>> create mode 100644 drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
>>>
>>>diff --git a/drivers/gpu/drm/nouveau/Makefile
>>>b/drivers/gpu/drm/nouveau/Makefile
>>>index 8b307e143632..11d9561d67c1 100644
>>>--- a/drivers/gpu/drm/nouveau/Makefile
>>>+++ b/drivers/gpu/drm/nouveau/Makefile
>>>@@ -26,6 +26,7 @@ nouveau-y += core/core/subdev.o
>>> nouveau-y += core/subdev/bar/base.o
>>> nouveau-y += core/subdev/bar/nv50.o
>>> nouveau-y += core/subdev/bar/nvc0.o
>>>+nouveau-y += core/subdev/bar/gk20a.o
>>> nouveau-y += core/subdev/bios/base.o
>>> nouveau-y += core/subdev/bios/bit.o
>>> nouveau-y += core/subdev/bios/boost.o
>>>diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
>>>b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
>>>index 2d1e97d4264f..a2b9ccc48f66 100644
>>>--- a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
>>>+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
>>>@@ -165,7 +165,7 @@ nve0_identify(struct nouveau_device *device)
>>> device->oclass[NVDEV_SUBDEV_IBUS ] =
>>>&gk20a_ibus_oclass;
>>> device->oclass[NVDEV_SUBDEV_INSTMEM] =
>>>nv50_instmem_oclass;
>>> device->oclass[NVDEV_SUBDEV_VM ] =
>>>&nvc0_vmmgr_oclass;
>>>- device->oclass[NVDEV_SUBDEV_BAR ] = &nvc0_bar_oclass;
>>>+ device->oclass[NVDEV_SUBDEV_BAR ] = &gk20a_bar_oclass;
>>> device->oclass[NVDEV_ENGINE_DMAOBJ ] =
>>>&nvd0_dmaeng_oclass;
>>> device->oclass[NVDEV_ENGINE_FIFO ] =
>>>gk20a_fifo_oclass;
>>> device->oclass[NVDEV_ENGINE_SW ] =
>>>nvc0_software_oclass;
>>>diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
>>>b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
>>>index 9002cbb6432b..be037fac534c 100644
>>>--- a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
>>>+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
>>>@@ -33,5 +33,6 @@ nouveau_bar(void *obj)
>>>
>>> extern struct nouveau_oclass nv50_bar_oclass;
>>> extern struct nouveau_oclass nvc0_bar_oclass;
>>>+extern struct nouveau_oclass gk20a_bar_oclass;
>>>
>>> #endif
>>>diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
>>>b/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
>>>new file mode 100644
>>>index 000000000000..bf877af9d3bd
>>>--- /dev/null
>>>+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
>>>@@ -0,0 +1,54 @@
>>>+/*
>>>+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
>>>+ *
>>>+ * Permission is hereby granted, free of charge, to any person
>>>obtaining
>>>a
>>>+ * copy of this software and associated documentation files (the
>>>"Software"),
>>>+ * to deal in the Software without restriction, including without
>>>limitation
>>>+ * the rights to use, copy, modify, merge, publish, distribute,
>>>sublicense,
>>>+ * and/or sell copies of the Software, and to permit persons to whom
>>>the
>>>+ * Software is furnished to do so, subject to the following conditions:
>>>+ *
>>>+ * The above copyright notice and this permission notice shall be
>>>included in
>>>+ * all copies or substantial portions of the Software.
>>>+ *
>>>+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
>>>EXPRESS OR
>>>+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
>>>MERCHANTABILITY,
>>>+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT
>>>SHALL
>>>+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
>>>OTHER
>>>+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
>>>ARISING
>>>+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
>>>+ * DEALINGS IN THE SOFTWARE.
>>>+ */
>>>+
>>>+#include <subdev/bar.h>
>>>+
>>>+#include "priv.h"
>>>+
>>>+int
>>>+gk20a_bar_ctor(struct nouveau_object *parent, struct nouveau_object
>>>*engine,
>>>+ struct nouveau_oclass *oclass, void *data, u32 size,
>>>+ struct nouveau_object **pobject)
>>>+{
>>>+ struct nouveau_bar *bar;
>>>+ int ret;
>>>+
>>>+ ret = nvc0_bar_ctor(parent, engine, oclass, data, size, pobject);
>>>+ if (ret)
>>>+ return ret;
>>>+
>>>+ bar = (struct nouveau_bar *)*pobject;
>>>+ bar->iomap_uncached = true;
>>>+
>>>+ return 0;
>>>+}
>>>+
>>>+struct nouveau_oclass
>>>+gk20a_bar_oclass = {
>>>+ .handle = NV_SUBDEV(BAR, 0xea),
>>>+ .ofuncs = &(struct nouveau_ofuncs) {
>>>+ .ctor = gk20a_bar_ctor,
>>>+ .dtor = nvc0_bar_dtor,
>>>+ .init = nvc0_bar_init,
>>>+ .fini = _nouveau_bar_fini,
>>>+ },
>>>+};
>>>diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
>>>b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
>>>index ca8139b9ab27..0a44459844e3 100644
>>>--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
>>>+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
>>>@@ -133,7 +133,7 @@ nvc0_bar_init_vm(struct nvc0_bar_priv *priv, struct
>>>nvc0_bar_priv_vm *bar_vm,
>>> return 0;
>>> }
>>>
>>>-static int
>>>+int
>>> nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object
>>>*engine,
>>> struct nouveau_oclass *oclass, void *data, u32 size,
>>> struct nouveau_object **pobject)
>>>@@ -169,7 +169,7 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct
>>>nouveau_object *engine,
>>> return 0;
>>> }
>>>
>>>-static void
>>>+void
>>> nvc0_bar_dtor(struct nouveau_object *object)
>>> {
>>> struct nvc0_bar_priv *priv = (void *)object;
>>>@@ -188,7 +188,7 @@ nvc0_bar_dtor(struct nouveau_object *object)
>>> nouveau_bar_destroy(&priv->base);
>>> }
>>>
>>>-static int
>>>+int
>>> nvc0_bar_init(struct nouveau_object *object)
>>> {
>>> struct nvc0_bar_priv *priv = (void *)object;
>>>diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
>>>b/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
>>>index ffad8f337ead..3ee8b1476d00 100644
>>>--- a/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
>>>+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
>>>@@ -23,4 +23,10 @@ int nouveau_bar_alloc(struct nouveau_bar *, struct
>>>nouveau_object *,
>>>
>>> void nv84_bar_flush(struct nouveau_bar *);
>>>
>>>+int nvc0_bar_ctor(struct nouveau_object *, struct nouveau_object *,
>>>+ struct nouveau_oclass *, void *, u32,
>>>+ struct nouveau_object **);
>>>+void nvc0_bar_dtor(struct nouveau_object *);
>>>+int nvc0_bar_init(struct nouveau_object *);
>>>+
>>> #endif
>>>--
>>>2.0.0
>>>
>>>--
>>>To unsubscribe from this list: send the line "unsubscribe linux-tegra"
>>>in
>>>the body of a message to majordomo@...r.kernel.org
>>>More majordomo info at http://vger.kernel.org/majordomo-info.html
>>
>> _______________________________________________
>> dri-devel mailing list
>> dri-devel@...ts.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/dri-devel
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists