Skip to contentSkip to conversation

drm/i915: Add scheduling priority to per-context parameters

RFC Superseded Dave Gordon 3 participants 5 messages Oct 1, 2015 – Oct 2, 2015

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover drm/i915: Add scheduling priority to per-context parameters

A proposal for an ioctl to allow a user process to alter its GPU scheduling priority. It's defined as a new per-context parameter, and each execbuffer submitted thereafter inherits its priority from the context. This allows a single process to associate different priorities with different contexts, perhaps being used for different purposes (e.g. updating the GUI vs. scanning SETI data for alien transmissions).

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#1 lore raw

RFC 1/1 drm/i915: Add scheduling priority to per-context parameters

The next use for the i915 get/set per-context parameters ioctl, ahead of the introduction of the forthcoming GPU scheduler.

---
 drivers/gpu/drm/i915/i915_drv.h         | 28 ++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_context.c | 17 +++++++++++++++++
 include/uapi/drm/i915_drm.h             |  1 +
 3 files changed, 46 insertions(+)
drivers/gpu/drm/i915/i915_drv.h +28 -0
index 279e258..104b711 100644
@@ -850,6 +850,33 @@ struct i915_ctx_hang_stats {
 	bool banned;
 };
 
+/*
+ * User-settable GFX scheduler priorities are on a scale of 1 (lowest
+ * priority) to 1023 (highest priority). The special value 0 means
+ * "let the system decide my priority automatically"; this is the
+ * default if the user process does not explicitly request a different
+ * priority. Any process may decrease its scheduling priority, but
+ * only a sufficiently-privileged process may increase it. However,
+ * it is always permissible to reset it to "system default", even if
+ * is currently lower than that. Thus, if the system-assigned default
+ * were, say, 256, a process could decrease it to 128, and then to 64.
+ * It could NOT then increase it to 128 again, but COULD request a
+ * priority of 0 -- which would actually reset it to 256, allowing
+ * the process to then request 128 again. (This avoids the issue with
+ * nice(2) priorities, namely that non-super-users can not increase
+ * scheduling priorities of their own processes, even if they were the
+ * ones that decreased the priorities in the first place).
+ */
+
+#define	i915_SCHED_PRIORITY_DEFAULT	0
+#define	i915_SCHED_PRIORITY_MIN		1
+#define	i915_SCHED_PRIORITY_MAX		1023
+
+struct i915_ctx_sched_info {
+	/* Scheduling priority */
+	unsigned long priority;
+};
+
 /* This must match up with the value previously used for execbuf2.rsvd1. */
 #define DEFAULT_CONTEXT_HANDLE 0
 
@@ -881,6 +908,7 @@ struct intel_context {
 	int flags;
 	struct drm_i915_file_private *file_priv;
 	struct i915_ctx_hang_stats hang_stats;
+	struct i915_ctx_sched_info sched_info;
 	struct i915_hw_ppgtt *ppgtt;
 
 	/* Legacy ring buffer submission */
drivers/gpu/drm/i915/i915_gem_context.c +17 -0
index 74aa0c9..2728a0d 100644
@@ -901,6 +901,9 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
 	case I915_CONTEXT_PARAM_NO_ZEROMAP:
 		args->value = ctx->flags & CONTEXT_NO_ZEROMAP;
 		break;
+	case I915_CONTEXT_PARAM_PRIORITY:
+		args->value = ctx->sched_info.priority;
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -938,6 +941,7 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 		else
 			ctx->hang_stats.ban_period_seconds = args->value;
 		break;
+
 	case I915_CONTEXT_PARAM_NO_ZEROMAP:
 		if (args->size) {
 			ret = -EINVAL;
@@ -946,6 +950,19 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 			ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0;
 		}
 		break;
+
+	case I915_CONTEXT_PARAM_PRIORITY:
+		if (args->size)
+			ret = -EINVAL;
+		else if (args->value > I915_SCHED_PRIORITY_MAX)
+			ret = -EINVAL;
+		else if (args->value > ctx->sched_info.priority &&
+			 !capable(CAP_SYS_ADMIN))
+			ret = -EPERM;
+		else
+			ctx->sched_info.priority = args->value;
+		break;
+
 	default:
 		ret = -EINVAL;
 		break;
include/uapi/drm/i915_drm.h +1 -0
index fd5aa47..a860263 100644
@@ -1126,6 +1126,7 @@ struct drm_i915_gem_context_param {
 	__u64 param;
 #define I915_CONTEXT_PARAM_BAN_PERIOD 0x1
 #define I915_CONTEXT_PARAM_NO_ZEROMAP 0x2
+#define I915_CONTEXT_PARAM_PRIORITY   0x3
 	__u64 value;
 };
 
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
config: x86_64-allyesconfig (attached as .config) reproduce:

Hi Dave,

[auto build test results on v4.3-rc3 -- if it's inappropriate base, please ignore]

config: x86_64-allyesconfig (attached as .config)
reproduce:

        git checkout 5be87551bfc8fabcb9cad4762a45c0f12d0c6cdf
        # save the attached .config to linux build tree
        make ARCH=x86_64 

All error/warnings (new ones prefixed by >>):

drivers/gpu/drm/i915/i915_gem_context.c: In function 'i915_gem_context_setparam_ioctl':

>> drivers/gpu/drm/i915/i915_gem_context.c:950:26: error: 'I915_SCHED_PRIORITY_MAX' undeclared (first use in this function)
      else if (args->value > I915_SCHED_PRIORITY_MAX)
                             ^

drivers/gpu/drm/i915/i915_gem_context.c:950:26: note: each undeclared identifier is reported only once for each function it appears in

vim +/I915_SCHED_PRIORITY_MAX +950 drivers/gpu/drm/i915/i915_gem_context.c

944 }
945 break;
946
947 case I915_CONTEXT_PARAM_PRIORITY:
948 if (args->size)
949 ret = -EINVAL;

 > 950			else if (args->value > I915_SCHED_PRIORITY_MAX)

951 ret = -EINVAL;
952 else if (args->value > ctx->sched_info.priority &&
953 !capable(CAP_SYS_ADMIN))

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw

RFC v2 1/1 drm/i915: Add scheduling priority to per-context parameters

In reply to Dave Gordon

Hmmm ... the email seems to have been damaged during composition :(
I probably shouldn't try to use vi(1) [where '~' means toggle-letter-case] over an ssh link [where '~' is an escape, of sorts] from another Linux machine inside a PuTTY terminal under Windows [where various keys send escape sequences containing '~'] :( Anyway, this version has the #defines as they actually appeared in the source,
i.e. starting with UPPERCASE 'I' and not lowercase 'i'!

The next use for the i915 get/set per-context parameters ioctl, ahead of the introduction of the forthcoming GPU scheduler.

---
 drivers/gpu/drm/i915/i915_drv.h         | 28 ++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_context.c | 17 +++++++++++++++++
 include/uapi/drm/i915_drm.h             |  1 +
 3 files changed, 46 insertions(+)
drivers/gpu/drm/i915/i915_drv.h +27 -0
index 279e258..104b711 100644
@@ -850,6 +850,33 @@ struct i915_ctx_hang_stats {
 	bool banned;
 };
 
+/*
+ * User-settable GFX scheduler priorities are on a scale of 1 (lowest
+ * priority) to 1023 (highest priority). The special value 0 means
+ * "let the system decide my priority automatically"; this is the
+ * default if the user process does not explicitly request a different
+ * priority. Any process may decrease its scheduling priority, but
+ * only a sufficiently-privileged process may increase it. However,
+ * it is always permissible to reset it to "system default", even if
+ * is currently lower than that. Thus, if the system-assigned default
+ * were, say, 256, a process could decrease it to 128, and then to 64.
+ * It could NOT then increase it to 128 again, but COULD request a
+ * priority of 0 -- which would actually reset it to 256, allowing
+ * the process to then request 128 again. (This avoids the issue with
+ * nice(2) priorities, namely that non-super-users can not increase
+ * scheduling priorities of their own processes, even if they were the
+ * ones that decreased the priorities in the first place).
+ */
+
+#define	I915_SCHED_PRIORITY_DEFAULT	0
+#define	I915_SCHED_PRIORITY_MIN		1
+#define	I915_SCHED_PRIORITY_MAX		1023
+
+struct i915_ctx_sched_info {
+	/* Scheduling priority */
+	unsigned long priority;
+};
+
 /* This must match up with the value previously used for 

execbuf2.rsvd1. */
#define DEFAULT_CONTEXT_HANDLE 0

@@ -881,6 +908,7 @@ struct intel_context {
int flags;
struct drm_i915_file_private *file_priv;
struct i915_ctx_hang_stats hang_stats;
+ struct i915_ctx_sched_info sched_info;
struct i915_hw_ppgtt *ppgtt;

/* Legacy ring buffer submission */

drivers/gpu/drm/i915/i915_gem_context.c +17 -0
index 74aa0c9..2728a0d 100644
@@ -901,6 +901,9 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
 	case I915_CONTEXT_PARAM_NO_ZEROMAP:
 		args->value = ctx->flags & CONTEXT_NO_ZEROMAP;
 		break;
+	case I915_CONTEXT_PARAM_PRIORITY:
+		args->value = ctx->sched_info.priority;
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -938,6 +941,7 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 		else
 			ctx->hang_stats.ban_period_seconds = args->value;
 		break;
+
 	case I915_CONTEXT_PARAM_NO_ZEROMAP:
 		if (args->size) {
 			ret = -EINVAL;
@@ -946,6 +950,19 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 			ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0;
 		}
 		break;
+
+	case I915_CONTEXT_PARAM_PRIORITY:
+		if (args->size)
+			ret = -EINVAL;
+		else if (args->value > I915_SCHED_PRIORITY_MAX)
+			ret = -EINVAL;
+		else if (args->value > ctx->sched_info.priority &&
+			 !capable(CAP_SYS_ADMIN))
+			ret = -EPERM;
+		else
+			ctx->sched_info.priority = args->value;
+		break;
+
 	default:
 		ret = -EINVAL;
 		break;
include/uapi/drm/i915_drm.h +1 -0
index fd5aa47..a860263 100644
@@ -1126,6 +1126,7 @@ struct drm_i915_gem_context_param {
 	__u64 param;
 #define I915_CONTEXT_PARAM_BAN_PERIOD 0x1
 #define I915_CONTEXT_PARAM_NO_ZEROMAP 0x2
+#define I915_CONTEXT_PARAM_PRIORITY   0x3
 	__u64 value;
 };
 
Signature · 8 lines
1.9.1



_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw
Quoted 34 lines · Dave Gordon

Hmmm ... the email seems to have been damaged during composition :( I probably shouldn't try to use vi(1) [where '~' means toggle-letter-case] over an ssh link [where '~' is an escape, of sorts] from another Linux machine inside a PuTTY terminal under Windows [where various keys send escape sequences containing '~'] :( Anyway, this
version has the #defines as they actually appeared in the source, i.e. starting with UPPERCASE 'I' and not lowercase 'i'!

The next use for the i915 get/set per-context parameters ioctl, ahead of the introduction of the forthcoming GPU scheduler.

---
 drivers/gpu/drm/i915/i915_drv.h         | 28 ++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_context.c | 17 +++++++++++++++++
 include/uapi/drm/i915_drm.h             |  1 +
 3 files changed, 46 insertions(+)
drivers/gpu/drm/i915/i915_drv.h +6 -0
index 279e258..104b711 100644
@@ -850,6 +850,33 @@ struct i915_ctx_hang_stats {
 	bool banned;
 };
 
+/*
+ * User-settable GFX scheduler priorities are on a scale of 1 (lowest
+ * priority) to 1023 (highest priority). The special value 0 means
+ * "let the system decide my priority automatically"; this is the
+ * default if the user process does not explicitly request a different
+ * priority. Any process may decrease its scheduling priority, but

drivers/gpu/drm/i915/i915_drv.h struct i915_ctx_hang_stats { · line 868

Quoted 10 lines
drivers/gpu/drm/i915/i915_drv.h +10 -0
+ * only a sufficiently-privileged process may increase it. However,
+ * it is always permissible to reset it to "system default", even if
+ * is currently lower than that. Thus, if the system-assigned default
+ * were, say, 256, a process could decrease it to 128, and then to 64.
+ * It could NOT then increase it to 128 again, but COULD request a
+ * priority of 0 -- which would actually reset it to 256, allowing
+ * the process to then request 128 again. (This avoids the issue with
+ * nice(2) priorities, namely that non-super-users can not increase
+ * scheduling priorities of their own processes, even if they were the
+ * ones that decreased the priorities in the first place).

I would prefer not to couple it in such a way. Have a continuous range -1024,1024 (default 0), but only allow a privileged process to request a positive priority value. So any process can set any negative/zero value at any time (thereby getting a small boost in priority at times), but only the select few can completely gazzump them by setting a positive value.

That is a little more intuitive from my perspective. Have you considered skipping the nice() step entirely and jump to a setpriority() model? -Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx