[PATCH 2/3] drm: Expand max DRM device number to full MINORBITS

Michał Winiarski michal.winiarski at intel.com
Wed Aug 17 23:05:59 UTC 2022


Even after getting rid of control nodes, we've only bumped max DRM
device number from 64 to 128.
That's not good enough for modern world where we have multi-GPU servers,
SR-IOV virtual functions and virtual devices used for testing.
Let's utilize full minor range for DRM devices.
To avoid regressing the existing userspace, we're still maintaining the
0-127 for primary, 128-255 for render. This numbering scheme is
continued for minors > 256 (256-383 for primary, 384-511 for render, and
so on).

Signed-off-by: Michał Winiarski <michal.winiarski at intel.com>
---
 drivers/gpu/drm/drm_drv.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index d81783f43452..0dab1ef8a98d 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -113,7 +113,7 @@ static int drm_minor_alloc(struct drm_device *dev, unsigned int type)
 {
 	struct drm_minor *minor;
 	unsigned long flags;
-	int r;
+	int r, start, end;
 
 	minor = drmm_kzalloc(dev, sizeof(*minor), GFP_KERNEL);
 	if (!minor)
@@ -122,15 +122,19 @@ static int drm_minor_alloc(struct drm_device *dev, unsigned int type)
 	minor->type = type;
 	minor->dev = dev;
 
-	idr_preload(GFP_KERNEL);
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	r = idr_alloc(&drm_minors_idr,
-		      NULL,
-		      128 * type,
-		      128 * (type + 1),
-		      GFP_NOWAIT);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-	idr_preload_end();
+	start = 128 * type;
+	end = 128 * (type + 1);
+
+	do {
+		idr_preload(GFP_KERNEL);
+		spin_lock_irqsave(&drm_minor_lock, flags);
+		r = idr_alloc(&drm_minors_idr, NULL, start, end, GFP_NOWAIT);
+		spin_unlock_irqrestore(&drm_minor_lock, flags);
+		idr_preload_end();
+
+		start += 256;
+		end += 256;
+	} while ((r == -ENOSPC) && end <= (1 << MINORBITS));
 
 	if (r < 0)
 		return r;
-- 
2.37.2



More information about the dri-devel mailing list