intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
* [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection
@ 2020-11-17 11:40 Tvrtko Ursulin
  2020-11-17 11:51 ` Chris Wilson
                   ` (2 more replies)
  0 siblings, 3 replies; 8+ messages in thread
From: Tvrtko Ursulin @ 2020-11-17 11:40 UTC (permalink / raw)
  To: igt-dev; +Cc: Intel-gfx

From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

-L and -D <device> on the command line.

With no device specified tool tries to find i915 discrete or integrated in
that order.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 benchmarks/gem_wsim.c | 61 +++++++++++++++++++++++++++++++++++--------
 1 file changed, 50 insertions(+), 11 deletions(-)

diff --git a/benchmarks/gem_wsim.c b/benchmarks/gem_wsim.c
index ecad4a8dcd22..f3567f7b9a0b 100644
--- a/benchmarks/gem_wsim.c
+++ b/benchmarks/gem_wsim.c
@@ -43,6 +43,7 @@
 #include <pthread.h>
 #include <math.h>
 
+#include "igt_device_scan.h"
 #include "intel_chipset.h"
 #include "intel_reg.h"
 #include "drm.h"
@@ -2520,7 +2521,9 @@ static void print_help(void)
 "                    clients.\n"
 "  -d                Sync between data dependencies in userspace.\n"
 "  -f <scale>        Scale factor for batch durations.\n"
-"  -F <scale>        Scale factor for delays."
+"  -F <scale>        Scale factor for delays.\n"
+"  -L                List GPUs.\n"
+"  -D <gpu>          One of the GPUs from -L.\n"
 	);
 }
 
@@ -2570,6 +2573,7 @@ add_workload_arg(struct w_arg *w_args, unsigned int nr_args, char *w_arg,
 
 int main(int argc, char **argv)
 {
+	bool list_devices_arg = false;
 	unsigned int repeat = 1;
 	unsigned int clients = 1;
 	unsigned int flags = 0;
@@ -2581,26 +2585,25 @@ int main(int argc, char **argv)
 	char *append_workload_arg = NULL;
 	struct w_arg *w_args = NULL;
 	int exitcode = EXIT_FAILURE;
+	struct igt_device_card card;
+	char *device_arg = NULL;
 	double scale_time = 1.0f;
 	double scale_dur = 1.0f;
 	int prio = 0;
 	double t;
 	int i, c, ret;
 
-	/*
-	 * Open the device via the low-level API so we can do the GPU quiesce
-	 * manually as close as possible in time to the start of the workload.
-	 * This minimizes the gap in engine utilization tracking when observed
-	 * via external tools like trace.pl.
-	 */
-	fd = __drm_open_driver_render(DRIVER_INTEL);
-	igt_require(fd);
-
 	master_prng = time(NULL);
 
 	while ((c = getopt(argc, argv,
-			   "hqvsSdc:r:w:W:a:p:I:f:F:")) != -1) {
+			   "LhqvsSdc:r:w:W:a:p:I:f:F:D:")) != -1) {
 		switch (c) {
+		case 'L':
+			list_devices_arg = true;
+			break;
+		case 'D':
+			device_arg = strdup(optarg);
+			break;
 		case 'W':
 			if (master_workload >= 0) {
 				wsim_err("Only one master workload can be given!\n");
@@ -2661,6 +2664,42 @@ int main(int argc, char **argv)
 		}
 	}
 
+	igt_devices_scan(false);
+
+	if (list_devices_arg) {
+		struct igt_devices_print_format fmt = {
+			.type = IGT_PRINT_USER,
+			.option = IGT_PRINT_DRM,
+		};
+
+		igt_devices_print(&fmt);
+
+		return EXIT_SUCCESS;
+	}
+
+	if (device_arg) {
+		ret = igt_device_card_match(device_arg, &card);
+		if (!ret) {
+			wsim_err("Requested device %s not found!\n",
+				 device_arg);
+			free(device_arg);
+
+			return EXIT_FAILURE;
+		}
+		free(device_arg);
+	} else {
+		ret = igt_device_find_first_i915_discrete_card(&card);
+		if (!ret)
+			ret = igt_device_find_integrated_card(&card);
+		if (!ret) {
+			wsim_err("No device filter specified and no discrete/integrated i915 devices found\n");
+			return EXIT_FAILURE;
+		}
+	}
+
+	fd = igt_open_card(&card);
+	igt_require(fd);
+
 	if (!nr_w_args) {
 		wsim_err("No workload descriptor(s)!\n");
 		goto err;
-- 
2.25.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection
  2020-11-17 11:40 [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection Tvrtko Ursulin
@ 2020-11-17 11:51 ` Chris Wilson
  2020-11-17 12:18 ` Chris Wilson
  2020-11-17 13:59 ` [Intel-gfx] [PATCH i-g-t v2] " Tvrtko Ursulin
  2 siblings, 0 replies; 8+ messages in thread
From: Chris Wilson @ 2020-11-17 11:51 UTC (permalink / raw)
  To: Tvrtko Ursulin, igt-dev; +Cc: Intel-gfx

Quoting Tvrtko Ursulin (2020-11-17 11:40:50)
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> -L and -D <device> on the command line.
> 
> With no device specified tool tries to find i915 discrete or integrated in
> that order.

I'm shocked you didn't add D.x to .wsim :)
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection
  2020-11-17 11:40 [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection Tvrtko Ursulin
  2020-11-17 11:51 ` Chris Wilson
@ 2020-11-17 12:18 ` Chris Wilson
  2020-11-17 12:33   ` Tvrtko Ursulin
  2020-11-17 13:59 ` [Intel-gfx] [PATCH i-g-t v2] " Tvrtko Ursulin
  2 siblings, 1 reply; 8+ messages in thread
From: Chris Wilson @ 2020-11-17 12:18 UTC (permalink / raw)
  To: Tvrtko Ursulin, igt-dev; +Cc: Intel-gfx

Quoting Tvrtko Ursulin (2020-11-17 11:40:50)
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> -L and -D <device> on the command line.
> 
> With no device specified tool tries to find i915 discrete or integrated in
> that order.

Ok, looks straightforward enough. (I still like having an idiot's guide
with examples of -L output and what to pass to -D).

> +       fd = igt_open_card(&card);
> +       igt_require(fd);

Does igt_require() still generate weirdness when used outside of igt
tests?

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection
  2020-11-17 12:18 ` Chris Wilson
@ 2020-11-17 12:33   ` Tvrtko Ursulin
  2020-11-17 12:37     ` Chris Wilson
  0 siblings, 1 reply; 8+ messages in thread
From: Tvrtko Ursulin @ 2020-11-17 12:33 UTC (permalink / raw)
  To: Chris Wilson, igt-dev; +Cc: Intel-gfx


On 17/11/2020 12:18, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2020-11-17 11:40:50)
>> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>>
>> -L and -D <device> on the command line.
>>
>> With no device specified tool tries to find i915 discrete or integrated in
>> that order.
> 
> Ok, looks straightforward enough. (I still like having an idiot's guide
> with examples of -L output and what to pass to -D).
> 
>> +       fd = igt_open_card(&card);
>> +       igt_require(fd);
> 
> Does igt_require() still generate weirdness when used outside of igt
> tests?

It appears to do nothing. But more importantly I don't know how I ended 
up with it here. Probably some copy & paste again.

> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

Can I convert to igt_assert and keep the r-b?

Regards,

Tvrtko

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection
  2020-11-17 12:33   ` Tvrtko Ursulin
@ 2020-11-17 12:37     ` Chris Wilson
  0 siblings, 0 replies; 8+ messages in thread
From: Chris Wilson @ 2020-11-17 12:37 UTC (permalink / raw)
  To: Tvrtko Ursulin, igt-dev; +Cc: Intel-gfx

Quoting Tvrtko Ursulin (2020-11-17 12:33:11)
> 
> On 17/11/2020 12:18, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2020-11-17 11:40:50)
> >> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> >>
> >> -L and -D <device> on the command line.
> >>
> >> With no device specified tool tries to find i915 discrete or integrated in
> >> that order.
> > 
> > Ok, looks straightforward enough. (I still like having an idiot's guide
> > with examples of -L output and what to pass to -D).
> > 
> >> +       fd = igt_open_card(&card);
> >> +       igt_require(fd);
> > 
> > Does igt_require() still generate weirdness when used outside of igt
> > tests?
> 
> It appears to do nothing. But more importantly I don't know how I ended 
> up with it here. Probably some copy & paste again.
> 
> > Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> Can I convert to igt_assert and keep the r-b?

Make it return sensibly with a wsim_err.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Intel-gfx] [PATCH i-g-t v2] gem_wsim: Implement device selection
  2020-11-17 11:40 [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection Tvrtko Ursulin
  2020-11-17 11:51 ` Chris Wilson
  2020-11-17 12:18 ` Chris Wilson
@ 2020-11-17 13:59 ` Tvrtko Ursulin
  2020-11-17 14:06   ` [Intel-gfx] [igt-dev] " Chris Wilson
  2 siblings, 1 reply; 8+ messages in thread
From: Tvrtko Ursulin @ 2020-11-17 13:59 UTC (permalink / raw)
  To: igt-dev; +Cc: Intel-gfx

From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

-L and -D <device> on the command line.

With no device specified tool tries to find i915 discrete or integrated in
that order.

v2:
 * Fix error handling and support render devices.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 benchmarks/gem_wsim.c | 75 ++++++++++++++++++++++++++++++++++++-------
 1 file changed, 64 insertions(+), 11 deletions(-)

diff --git a/benchmarks/gem_wsim.c b/benchmarks/gem_wsim.c
index ecad4a8dcd22..1c0c591ed8af 100644
--- a/benchmarks/gem_wsim.c
+++ b/benchmarks/gem_wsim.c
@@ -43,6 +43,7 @@
 #include <pthread.h>
 #include <math.h>
 
+#include "igt_device_scan.h"
 #include "intel_chipset.h"
 #include "intel_reg.h"
 #include "drm.h"
@@ -2520,7 +2521,9 @@ static void print_help(void)
 "                    clients.\n"
 "  -d                Sync between data dependencies in userspace.\n"
 "  -f <scale>        Scale factor for batch durations.\n"
-"  -F <scale>        Scale factor for delays."
+"  -F <scale>        Scale factor for delays.\n"
+"  -L                List GPUs.\n"
+"  -D <gpu>          One of the GPUs from -L.\n"
 	);
 }
 
@@ -2570,6 +2573,8 @@ add_workload_arg(struct w_arg *w_args, unsigned int nr_args, char *w_arg,
 
 int main(int argc, char **argv)
 {
+	struct igt_device_card card = { };
+	bool list_devices_arg = false;
 	unsigned int repeat = 1;
 	unsigned int clients = 1;
 	unsigned int flags = 0;
@@ -2581,26 +2586,25 @@ int main(int argc, char **argv)
 	char *append_workload_arg = NULL;
 	struct w_arg *w_args = NULL;
 	int exitcode = EXIT_FAILURE;
+	char *device_arg = NULL;
 	double scale_time = 1.0f;
 	double scale_dur = 1.0f;
 	int prio = 0;
 	double t;
 	int i, c, ret;
-
-	/*
-	 * Open the device via the low-level API so we can do the GPU quiesce
-	 * manually as close as possible in time to the start of the workload.
-	 * This minimizes the gap in engine utilization tracking when observed
-	 * via external tools like trace.pl.
-	 */
-	fd = __drm_open_driver_render(DRIVER_INTEL);
-	igt_require(fd);
+	char *drm_dev;
 
 	master_prng = time(NULL);
 
 	while ((c = getopt(argc, argv,
-			   "hqvsSdc:r:w:W:a:p:I:f:F:")) != -1) {
+			   "LhqvsSdc:r:w:W:a:p:I:f:F:D:")) != -1) {
 		switch (c) {
+		case 'L':
+			list_devices_arg = true;
+			break;
+		case 'D':
+			device_arg = strdup(optarg);
+			break;
 		case 'W':
 			if (master_workload >= 0) {
 				wsim_err("Only one master workload can be given!\n");
@@ -2661,6 +2665,55 @@ int main(int argc, char **argv)
 		}
 	}
 
+	igt_devices_scan(false);
+
+	if (list_devices_arg) {
+		struct igt_devices_print_format fmt = {
+			.type = IGT_PRINT_USER,
+			.option = IGT_PRINT_DRM,
+		};
+
+		igt_devices_print(&fmt);
+		return EXIT_SUCCESS;
+	}
+
+	if (device_arg) {
+		ret = igt_device_card_match(device_arg, &card);
+		if (!ret) {
+			wsim_err("Requested device %s not found!\n",
+				 device_arg);
+			free(device_arg);
+			return EXIT_FAILURE;
+		}
+		free(device_arg);
+	} else {
+		ret = igt_device_find_first_i915_discrete_card(&card);
+		if (!ret)
+			ret = igt_device_find_integrated_card(&card);
+		if (!ret) {
+			wsim_err("No device filter specified and no i915 devices found!\n");
+			return EXIT_FAILURE;
+		}
+	}
+
+	if (strlen(card.card)) {
+		drm_dev = card.card;
+	} else if (strlen(card.render)) {
+		drm_dev = card.render;
+	} else {
+		wsim_err("Failed to detect device!\n");
+		return EXIT_FAILURE;
+	}
+
+	fd = open(drm_dev, O_RDWR);
+	if (fd < 0) {
+		wsim_err("Failed to open '%s'! (%s)\n",
+			 drm_dev, strerror(errno));
+		return EXIT_FAILURE;
+	}
+	if (verbose > 1)
+		printf("Using device %s\n", drm_dev);
+
 	if (!nr_w_args) {
 		wsim_err("No workload descriptor(s)!\n");
 		goto err;
-- 
2.25.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [igt-dev] [PATCH i-g-t v2] gem_wsim: Implement device selection
  2020-11-17 13:59 ` [Intel-gfx] [PATCH i-g-t v2] " Tvrtko Ursulin
@ 2020-11-17 14:06   ` Chris Wilson
  2020-11-18 11:56     ` Tvrtko Ursulin
  0 siblings, 1 reply; 8+ messages in thread
From: Chris Wilson @ 2020-11-17 14:06 UTC (permalink / raw)
  To: Tvrtko Ursulin, igt-dev; +Cc: Intel-gfx

Quoting Tvrtko Ursulin (2020-11-17 13:59:18)
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> -L and -D <device> on the command line.
> 
> With no device specified tool tries to find i915 discrete or integrated in
> that order.
> 
> v2:
>  * Fix error handling and support render devices.
> 
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

Hmm, optarg should be a pointer into the argv, so is the copy necessary?
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [Intel-gfx] [igt-dev] [PATCH i-g-t v2] gem_wsim: Implement device selection
  2020-11-17 14:06   ` [Intel-gfx] [igt-dev] " Chris Wilson
@ 2020-11-18 11:56     ` Tvrtko Ursulin
  0 siblings, 0 replies; 8+ messages in thread
From: Tvrtko Ursulin @ 2020-11-18 11:56 UTC (permalink / raw)
  To: Chris Wilson, igt-dev; +Cc: Intel-gfx


On 17/11/2020 14:06, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2020-11-17 13:59:18)
>> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>>
>> -L and -D <device> on the command line.
>>
>> With no device specified tool tries to find i915 discrete or integrated in
>> that order.
>>
>> v2:
>>   * Fix error handling and support render devices.
>>
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> Hmm, optarg should be a pointer into the argv, so is the copy necessary?

I don't remember if I tried, and I don't see that getopt(3) explains 
optarg from that angle, so probably safest to keep as is. Too lazy to 
try.. Or maybe presence of strdup suggests I did already try and that 
optarg is indeed temporary copy managed by getopt.

Regards,

Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2020-11-18 11:56 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-17 11:40 [Intel-gfx] [PATCH i-g-t] gem_wsim: Implement device selection Tvrtko Ursulin
2020-11-17 11:51 ` Chris Wilson
2020-11-17 12:18 ` Chris Wilson
2020-11-17 12:33   ` Tvrtko Ursulin
2020-11-17 12:37     ` Chris Wilson
2020-11-17 13:59 ` [Intel-gfx] [PATCH i-g-t v2] " Tvrtko Ursulin
2020-11-17 14:06   ` [Intel-gfx] [igt-dev] " Chris Wilson
2020-11-18 11:56     ` Tvrtko Ursulin

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).