From: Ben Widawsky <bwidawsk@gmail.com>
To: intel-gfx@lists.freedesktop.org
Cc: benjamin.widawsky@linux.intel.com
Subject: [RFC] drm/i915: context support unit test-r2
Date: Fri, 14 Jan 2011 23:28:28 -0800 [thread overview]
Message-ID: <20110115072828.GA20773@lundgren.kumite> (raw)
[-- Attachment #1: Type: text/plain, Size: 1122 bytes --]
I've been sitting on this for over a week now. I figure I may as well
post this since I've been working on the implementation in libdrm and
the driver for the last week, although because of a bunch of
bureaucratic distractions this week, it amounted to only a day's worth
of work :P.
The changes basically implemented the slot mechanism that Daniel and
Kristian recommended. I'll admit that even this basic unit test has
changed a few times in the last few days, but it's pretty simple, and
shows the basic idea.
Most of the implementation details of the ctx related functions are
missing, but I think the names are descriptive enough. Let me know if
more details are needed, and I can add them.
I'd like to keep the implementation details such as, how I'll be pinning
the HW context memory, out of the API discussion - which is the
direction that I'd like this thread to move towards.
As a final note, the test copied the bad_blit test, which was the
easiest for me to reuse. I think blit is a really poor example for
context support, but it's also much easier for non-graphics people like
myself to understand.
Ben
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: gem_ctx_basic.c --]
[-- Type: text/x-c; charset=unknown-8bit, Size: 4606 bytes --]
/*
* Copyright © 2011 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include "drmtest.h"
#include "intel_batchbuffer.h"
#include "intel_gpu_tools.h"
#include "drm.h"
#include "i915_drm.h"
#include "intel_bufmgr.h"
#include "intel_context.h"
static drm_intel_bufmgr *bufmgr;
static drm_intel_ctx *ctx;
struct intel_batchbuffer *batch;
/* Would normally malloc this */
#define MAX_BOS_PER_BATCH 2
static drm_intel_ctx_flags ctx_flags[MAX_BOS_PER_BATCH];
#define RSVD_BUF_SLOT 0
#define SRC_BUF_SLOT 1
#define DEST_BUF_SLOT 2
/* XXX: We'll turn on ppgtt, but it's real usage won't get until gen7 */
#undef USE_PPGTT
static void
fill_ctx_bo_data(int which, drm_intel_bo *bo, int slot, unsigned long addr)
{
if (addr != 0) {
printf("Warning, PPGTT/user controlled relocations are not yet"
" supported\n");
}
/* XXX: reference to the bo handle */
ctx_flags[which].bo = bo;
/* XXX: which slot the dest buffer refers to in the context */
ctx_flags[which].slot = slot;
/* XXX: place to put the buffer (ppgtt only) */
ctx_flags[which].addr = addr;
/* XXX: domains */
ctx_flags[which].read_domain = I915_GEM_DOMAIN_RENDER;
ctx_flags[which].write_domain = 0;
/* XXX: what else? */
//drm_intel_ctx_set_cmd(&ctx_flags[which], CTX_BUF_ASSOCIATE);
ctx_flags[which].command = CTX_BUF_ASSOCIATE;
}
static void
blit(drm_intel_bo *src_bo, drm_intel_bo *dest_bo)
{
uint32_t src_pitch = 512, dst_pitch = 512;
uint32_t cmd_bits = 0;
/* XXX: populate context info for the src and dest buffers */
fill_ctx_bo_data(0, dest_bo, DEST_BUF_SLOT, 0x1000);
fill_ctx_bo_data(1, src_bo, SRC_BUF_SLOT, 0x8000);
drm_intel_ctx_set_flag_count(ctx, 2);
BEGIN_BATCH(8);
OUT_BATCH(XY_SRC_COPY_BLT_CMD |
XY_SRC_COPY_BLT_WRITE_ALPHA |
XY_SRC_COPY_BLT_WRITE_RGB |
cmd_bits);
OUT_BATCH((3 << 24) | /* 32 bits */
(0xcc << 16) | /* copy ROP */
dst_pitch);
OUT_BATCH(0); /* dst x1,y1 */
OUT_BATCH((64 << 16) | 64); /* 64x64 blit */
#ifdef USE_PPGTT
OUT_BATCH(ctx_flags[0].addr);
#else
OUT_RELOC(dest_bo, I915_GEM_DOMAIN_RENDER, 0, 0);
#endif
OUT_BATCH(0); /* src x1,y1 */
OUT_BATCH(src_pitch);
#ifdef USE_PPGTT
OUT_BATCH(ctx_flags[1].addr);
#else
OUT_RELOC(src_bo, I915_GEM_DOMAIN_RENDER, 0, 0);
#endif
ADVANCE_BATCH();
/* XXX: new flush with context data, where ctx has a pointer
* to the array of information per buffer object
*/
intel_batchbuffer_flush_with_ctx(batch, ctx);
}
int main(int argc, char **argv)
{
drm_intel_bo *src, *dest;
int fd;
const unsigned int aperture_size = 64 * 1024 * 1024;
/* XXX rsvd slot + src slot + dest slot */
const int ctx_bo_slots = 3;
fd = drm_open_any();
bufmgr = drm_intel_bufmgr_gem_init(fd, 4096);
drm_intel_bufmgr_set_debug(bufmgr, 1);
drm_intel_bufmgr_gem_enable_reuse(bufmgr);
/* XXX: Create a new context for our work 64MB aperture, 3 slots*/
ctx = drm_intel_ctx_gem_init(bufmgr, aperture_size, ctx_bo_slots);
if (ctx == NULL)
return -1;
/* XXX: associate the flags per bo with the context */
drm_intel_ctx_set_flags(ctx, ctx_flags);
batch = intel_batchbuffer_alloc(bufmgr, intel_get_drm_devid(fd));
src = drm_intel_bo_alloc(bufmgr, "src", 128 * 128, 4096);
dest = drm_intel_bo_alloc(bufmgr, "dest", 128 * 128, 4096);
blit(src, dest);
intel_batchbuffer_free(batch);
drm_intel_ctx_gem_destroy(ctx);
drm_intel_bufmgr_destroy(bufmgr);
close(fd);
return 0;
}
[-- Attachment #3: Type: text/plain, Size: 159 bytes --]
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
next reply other threads:[~2011-01-15 7:28 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-01-15 7:28 Ben Widawsky [this message]
2011-01-19 0:19 ` [RFC] drm/i915: context support unit test-r2 Ben Widawsky
2011-01-25 0:18 ` Ben Widawsky
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=20110115072828.GA20773@lundgren.kumite \
--to=bwidawsk@gmail.com \
--cc=benjamin.widawsky@linux.intel.com \
--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