From: David Herrmann <dh.herrmann@gmail.com>
To: dri-devel@lists.freedesktop.org
Cc: Ingo Molnar <mingo@kernel.org>,
linux-fbdev@vger.kernel.org, Dave Airlie <airlied@gmail.com>,
Daniel Vetter <daniel.vetter@ffwll.ch>,
Tomi Valkeinen <tomi.valkeinen@ti.com>,
linux-kernel@vger.kernel.org, Tom Gundersen <teg@jklm.no>,
David Herrmann <dh.herrmann@gmail.com>
Subject: [PATCH 08/11] drm/i915: remove sysfbs early
Date: Thu, 23 Jan 2014 14:15:00 +0000 [thread overview]
Message-ID: <1390486503-1504-9-git-send-email-dh.herrmann@gmail.com> (raw)
In-Reply-To: <1390486503-1504-1-git-send-email-dh.herrmann@gmail.com>
Once we allow DRM drivers for system-framebuffers, we need to evict such
devices *before* probing the real driver. A simple call to sysfb_claim()
does this and remove_conflicting_framebuffers() implicitly calls this.
However, it causes the sysfb device to be unloaded and thus locks
drm_global_mutex. remove_conflicting_framebuffers() must be called from
outside any ->load() callback to avoid a dead-lock.
All other DRM drivers call this right before probing the pci-device, which
is fine. For i915 we need to figure out the apertures before we can evict
fw-framebuffers, though. This turns out to be not as easy as you might
think, so lets just evict all sysfbs for now before loading i915.
A proper fix would be to make DRM code allow parallel device probing.
That's not going to happen soon, so be safe and make i915 evict all
sysfbs.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
---
drivers/gpu/drm/i915/i915_drv.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 43245b3..ceb875a 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -838,6 +838,12 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
driver.driver_features &= ~(DRIVER_USE_AGP);
+ /* We cannot call remove_conflicting_framebuffers() here as we cannot
+ * easily figure out the apertures here. So lets just remove all
+ * system-framebuffers early so we don't deadlock later when calling it
+ * with drm_global_mutex held. */
+ sysfb_claim(NULL, SYSFB_CLAIM_ALL);
+
return drm_get_pci_dev(pdev, ent, &driver);
}
--
1.8.5.3
WARNING: multiple messages have this Message-ID (diff)
From: David Herrmann <dh.herrmann@gmail.com>
To: dri-devel@lists.freedesktop.org
Cc: Ingo Molnar <mingo@kernel.org>,
linux-fbdev@vger.kernel.org, Dave Airlie <airlied@gmail.com>,
Daniel Vetter <daniel.vetter@ffwll.ch>,
Tomi Valkeinen <tomi.valkeinen@ti.com>,
linux-kernel@vger.kernel.org, Tom Gundersen <teg@jklm.no>,
David Herrmann <dh.herrmann@gmail.com>
Subject: [PATCH 08/11] drm/i915: remove sysfbs early
Date: Thu, 23 Jan 2014 15:15:00 +0100 [thread overview]
Message-ID: <1390486503-1504-9-git-send-email-dh.herrmann@gmail.com> (raw)
In-Reply-To: <1390486503-1504-1-git-send-email-dh.herrmann@gmail.com>
Once we allow DRM drivers for system-framebuffers, we need to evict such
devices *before* probing the real driver. A simple call to sysfb_claim()
does this and remove_conflicting_framebuffers() implicitly calls this.
However, it causes the sysfb device to be unloaded and thus locks
drm_global_mutex. remove_conflicting_framebuffers() must be called from
outside any ->load() callback to avoid a dead-lock.
All other DRM drivers call this right before probing the pci-device, which
is fine. For i915 we need to figure out the apertures before we can evict
fw-framebuffers, though. This turns out to be not as easy as you might
think, so lets just evict all sysfbs for now before loading i915.
A proper fix would be to make DRM code allow parallel device probing.
That's not going to happen soon, so be safe and make i915 evict all
sysfbs.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
---
drivers/gpu/drm/i915/i915_drv.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 43245b3..ceb875a 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -838,6 +838,12 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
driver.driver_features &= ~(DRIVER_USE_AGP);
+ /* We cannot call remove_conflicting_framebuffers() here as we cannot
+ * easily figure out the apertures here. So lets just remove all
+ * system-framebuffers early so we don't deadlock later when calling it
+ * with drm_global_mutex held. */
+ sysfb_claim(NULL, SYSFB_CLAIM_ALL);
+
return drm_get_pci_dev(pdev, ent, &driver);
}
--
1.8.5.3
next prev parent reply other threads:[~2014-01-23 14:15 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-23 14:14 [PATCH 00/11] SimpleDRM & Sysfb David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 01/11] x86: sysfb: fool-proof CONFIG_X86_SYSFB David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 02/11] x86: sysfb: remove sysfb when probing real hw David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 16:51 ` Ingo Molnar
2014-01-23 16:51 ` Ingo Molnar
2014-01-23 17:07 ` David Herrmann
2014-01-23 17:07 ` David Herrmann
2014-01-23 17:07 ` David Herrmann
2014-01-23 17:14 ` Ingo Molnar
2014-01-23 17:14 ` Ingo Molnar
2014-01-23 19:09 ` David Herrmann
2014-01-23 19:09 ` David Herrmann
2014-01-23 19:09 ` David Herrmann
2014-01-24 10:16 ` Ingo Molnar
2014-01-24 10:16 ` Ingo Molnar
2014-01-23 14:14 ` [PATCH 03/11] fbdev: efifb: add dev->remove() callback David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 04/11] fbdev: vesafb: " David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 05/11] x86: sysfb: store apertures in simplefb platform-data David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 06/11] video: sysfb: add generic firmware-fb interface David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:14 ` [PATCH 07/11] drm: mgag200: remove redundant fbdev removal David Herrmann
2014-01-23 14:14 ` David Herrmann
2014-01-23 14:15 ` David Herrmann [this message]
2014-01-23 14:15 ` [PATCH 08/11] drm/i915: remove sysfbs early David Herrmann
2014-01-23 14:15 ` [PATCH 09/11] drm: add SimpleDRM driver David Herrmann
2014-01-23 14:15 ` David Herrmann
2014-01-23 14:15 ` [PATCH 10/11] drm: simpledrm: add fbdev fallback support David Herrmann
2014-01-23 14:15 ` David Herrmann
2014-01-23 14:15 ` David Herrmann
2014-01-23 14:15 ` [PATCH 11/11] x86/sysfb: allow sysfb+simpledrm combination David Herrmann
2014-01-23 14:15 ` David Herrmann
2014-01-23 14:15 ` David Herrmann
2014-01-27 22:18 ` [PATCH 00/11] SimpleDRM & Sysfb David Herrmann
2014-01-27 22:18 ` David Herrmann
2014-01-27 22:18 ` David Herrmann
2014-02-21 9:56 ` Thierry Reding
2014-02-21 9:56 ` Thierry Reding
2014-02-21 9:56 ` Thierry Reding
2014-03-03 10:12 ` Tomi Valkeinen
2014-03-03 10:12 ` Tomi Valkeinen
2014-03-03 10:12 ` Tomi Valkeinen
2014-03-03 10:29 ` David Herrmann
2014-03-03 10:29 ` David Herrmann
2014-03-03 10:29 ` David Herrmann
2014-03-03 10:45 ` Tomi Valkeinen
2014-03-03 10:45 ` Tomi Valkeinen
2014-03-03 11:09 ` David Herrmann
2014-03-03 11:09 ` David Herrmann
2014-03-03 11:09 ` David Herrmann
2014-03-03 11:22 ` Tomi Valkeinen
2014-03-03 11:22 ` Tomi Valkeinen
2014-03-06 12:16 ` David Herrmann
2014-03-06 12:16 ` David Herrmann
2014-03-06 12:16 ` David Herrmann
2014-03-07 12:44 ` Tomi Valkeinen
2014-03-07 12:44 ` Tomi Valkeinen
2014-03-07 13:05 ` David Herrmann
2014-03-07 13:05 ` David Herrmann
2014-03-07 13:05 ` David Herrmann
2014-03-07 13:52 ` Tomi Valkeinen
2014-03-07 13:52 ` Tomi Valkeinen
2014-03-07 14:06 ` David Herrmann
2014-03-07 14:06 ` David Herrmann
2014-03-07 14:06 ` David Herrmann
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=1390486503-1504-9-git-send-email-dh.herrmann@gmail.com \
--to=dh.herrmann@gmail.com \
--cc=airlied@gmail.com \
--cc=daniel.vetter@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=linux-fbdev@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=teg@jklm.no \
--cc=tomi.valkeinen@ti.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.