public inbox for igt-dev@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Ramalingam C <ramalingam.c@intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: igt-dev@lists.freedesktop.org, intel-gfx@lists.freedesktop.org
Subject: Re: [igt-dev] [PATCH i-g-t] tests/dumb_buffer: Try to compute the largest possible dumb buffer
Date: Mon, 27 Jan 2020 12:11:10 +0530	[thread overview]
Message-ID: <20200127064110.GB1840@intel.com> (raw)
In-Reply-To: <20200123122711.1390696-1-chris@chris-wilson.co.uk>

On 2020-01-23 at 12:27:11 +0000, Chris Wilson wrote:
> For our threaded clear test, we want to create as many buffers as can
> fit into the available memory. However, since we don't know the size of
> that available memory, it can be easy to create too large or too many
> buffers, and waste our time testing the unobtainable. Instead, try
> probing the available space by trying to create the largest dumb buffer
> we possibly can, and assume that is the equivalent to the available
> space.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Ramalingam C <ramalingam.c@intel.com>
> ---
>  tests/dumb_buffer.c | 68 +++++++++++++++++++++++++++++++++++++++------
>  1 file changed, 59 insertions(+), 9 deletions(-)
> 
> diff --git a/tests/dumb_buffer.c b/tests/dumb_buffer.c
> index 3d2dc9966..c1e7b4d3b 100644
> --- a/tests/dumb_buffer.c
> +++ b/tests/dumb_buffer.c
> @@ -31,18 +31,18 @@
>   * combinations are rejected.
>   */
>  
> -#include <stdlib.h>
> -#include <sys/ioctl.h>
> -#include <stdio.h>
> -#include <string.h>
> +#include <errno.h>
>  #include <fcntl.h>
>  #include <inttypes.h>
> -#include <errno.h>
> -#include <sys/stat.h>
> -#include <sys/time.h>
> -#include <getopt.h>
>  #include <pthread.h>
> +#include <signal.h>
>  #include <stdatomic.h>
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <string.h>
> +#include <sys/ioctl.h>
> +#include <sys/stat.h>
> +#include <sys/time.h>
>  
>  #include <drm.h>
>  
> @@ -325,12 +325,62 @@ static void *thread_clear(void *data)
>  	return (void *)(uintptr_t)checked;
>  }
>  
> +static jmp_buf sigjmp;
> +
> +static void sigprobe(int sig)
> +{
> +	longjmp(sigjmp, sig);
> +}
> +
> +static uint64_t estimate_largest_dumb_buffer(int fd)
> +{
> +	sighandler_t old_sigbus = signal(SIGBUS, sigprobe);
> +	sighandler_t old_sigsegv = signal(SIGSEGV, sigprobe);
> +	struct drm_mode_create_dumb create = {
> +		.bpp = 32,
> +		.width = 1 << 20, /* in pixels */
> +		.height = 1, /* in rows */
> +	};
> +	volatile uint64_t largest = 0;
> +	char * volatile ptr = NULL;
> +
> +	if (setjmp(sigjmp)) {
> +		if (ptr)
> +			munmap(ptr, create.size);
> +
> +		signal(SIGBUS, old_sigbus);
> +		signal(SIGSEGV, old_sigsegv);
> +
> +		igt_info("Largest dumb buffer sucessfully created: %'"PRIu64" bytes\n",
> +			 largest);
> +		return largest / PAGE_SIZE;
> +	}
> +
> +	for (create.height = 1; create.height; create.height *= 2) {
> +		if (__dumb_create(fd, &create))
> +			longjmp(sigjmp, SIGABRT);
> +
> +		ptr = __dumb_map(fd, create.handle, create.size, PROT_READ);
> +		dumb_destroy(fd, create.handle);
> +		if (!ptr)
> +			longjmp(sigjmp, SIGABRT);
> +
> +		if (!*ptr)
> +			largest = create.size;
> +
> +		munmap(ptr, create.size);
> +		ptr = NULL;
> +	}
> +
> +	longjmp(sigjmp, SIGABRT);
> +}
> +
>  static void always_clear(int fd, int timeout)
>  {
>  	struct thread_clear arg = {
>  		.fd = fd,
>  		.timeout = timeout,
> -		.max = intel_get_avail_ram_mb() << (20 - 12), /* in pages */
> +		.max = estimate_largest_dumb_buffer(fd), /* in pages */
LGTM. I will test with other required LMEM changes.

Reviewed-by: Ramalingam C <ramalingam.c@intel.com>
>  	};
>  	const int ncpus = sysconf(_SC_NPROCESSORS_ONLN);
>  	unsigned long checked;
> -- 
> 2.25.0
> 
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev

      parent reply	other threads:[~2020-01-27  6:41 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-23 12:27 [igt-dev] [PATCH i-g-t] tests/dumb_buffer: Try to compute the largest possible dumb buffer Chris Wilson
2020-01-23 13:37 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2020-01-24  6:48 ` [igt-dev] ✗ GitLab.Pipeline: warning " Patchwork
2020-01-24 18:20 ` [igt-dev] ✗ Fi.CI.IGT: failure " Patchwork
2020-01-27  6:41 ` Ramalingam C [this message]

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=20200127064110.GB1840@intel.com \
    --to=ramalingam.c@intel.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox