linux/drivers/video/via/accel.c

548 lines
14 KiB
C
Raw Normal View History

/*
* Copyright 1998-2008 VIA Technologies, Inc. All Rights Reserved.
* Copyright 2001-2008 S3 Graphics, Inc. All Rights Reserved.
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License as published by the Free Software Foundation;
* either version 2, or (at your option) any later version.
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTIES OR REPRESENTATIONS; without even
* the implied warranty of MERCHANTABILITY or FITNESS FOR
* A PARTICULAR PURPOSE.See the GNU General Public License
* for more details.
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <linux/via-core.h>
#include "global.h"
/*
* Figure out an appropriate bytes-per-pixel setting.
*/
static int viafb_set_bpp(void __iomem *engine, u8 bpp)
{
u32 gemode;
/* Preserve the reserved bits */
/* Lowest 2 bits to zero gives us no rotation */
gemode = readl(engine + VIA_REG_GEMODE) & 0xfffffcfc;
switch (bpp) {
case 8:
gemode |= VIA_GEM_8bpp;
break;
case 16:
gemode |= VIA_GEM_16bpp;
break;
case 32:
gemode |= VIA_GEM_32bpp;
break;
default:
printk(KERN_WARNING "viafb_set_bpp: Unsupported bpp %d\n", bpp);
return -EINVAL;
}
writel(gemode, engine + VIA_REG_GEMODE);
return 0;
}
static int hw_bitblt_1(void __iomem *engine, u8 op, u32 width, u32 height,
u8 dst_bpp, u32 dst_addr, u32 dst_pitch, u32 dst_x, u32 dst_y,
u32 *src_mem, u32 src_addr, u32 src_pitch, u32 src_x, u32 src_y,
u32 fg_color, u32 bg_color, u8 fill_rop)
{
u32 ge_cmd = 0, tmp, i;
int ret;
if (!op || op > 3) {
printk(KERN_WARNING "hw_bitblt_1: Invalid operation: %d\n", op);
return -EINVAL;
}
if (op != VIA_BITBLT_FILL && !src_mem && src_addr == dst_addr) {
if (src_x < dst_x) {
ge_cmd |= 0x00008000;
src_x += width - 1;
dst_x += width - 1;
}
if (src_y < dst_y) {
ge_cmd |= 0x00004000;
src_y += height - 1;
dst_y += height - 1;
}
}
if (op == VIA_BITBLT_FILL) {
switch (fill_rop) {
case 0x00: /* blackness */
case 0x5A: /* pattern inversion */
case 0xF0: /* pattern copy */
case 0xFF: /* whiteness */
break;
default:
printk(KERN_WARNING "hw_bitblt_1: Invalid fill rop: "
"%u\n", fill_rop);
return -EINVAL;
}
}
ret = viafb_set_bpp(engine, dst_bpp);
if (ret)
return ret;
if (op != VIA_BITBLT_FILL) {
if (src_x & (op == VIA_BITBLT_MONO ? 0xFFFF8000 : 0xFFFFF000)
|| src_y & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported source "
"x/y %d %d\n", src_x, src_y);
return -EINVAL;
}
tmp = src_x | (src_y << 16);
writel(tmp, engine + 0x08);
}
if (dst_x & 0xFFFFF000 || dst_y & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported destination x/y "
"%d %d\n", dst_x, dst_y);
return -EINVAL;
}
tmp = dst_x | (dst_y << 16);
writel(tmp, engine + 0x0C);
if ((width - 1) & 0xFFFFF000 || (height - 1) & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported width/height "
"%d %d\n", width, height);
return -EINVAL;
}
tmp = (width - 1) | ((height - 1) << 16);
writel(tmp, engine + 0x10);
if (op != VIA_BITBLT_COLOR)
writel(fg_color, engine + 0x18);
if (op == VIA_BITBLT_MONO)
writel(bg_color, engine + 0x1C);
if (op != VIA_BITBLT_FILL) {
tmp = src_mem ? 0 : src_addr;
if (dst_addr & 0xE0000007) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported source "
"address %X\n", tmp);
return -EINVAL;
}
tmp >>= 3;
writel(tmp, engine + 0x30);
}
if (dst_addr & 0xE0000007) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported destination "
"address %X\n", dst_addr);
return -EINVAL;
}
tmp = dst_addr >> 3;
writel(tmp, engine + 0x34);
if (op == VIA_BITBLT_FILL)
tmp = 0;
else
tmp = src_pitch;
if (tmp & 0xFFFFC007 || dst_pitch & 0xFFFFC007) {
printk(KERN_WARNING "hw_bitblt_1: Unsupported pitch %X %X\n",
tmp, dst_pitch);
return -EINVAL;
}
tmp = VIA_PITCH_ENABLE | (tmp >> 3) | (dst_pitch << (16 - 3));
writel(tmp, engine + 0x38);
if (op == VIA_BITBLT_FILL)
ge_cmd |= fill_rop << 24 | 0x00002000 | 0x00000001;
else {
ge_cmd |= 0xCC000000; /* ROP=SRCCOPY */
if (src_mem)
ge_cmd |= 0x00000040;
if (op == VIA_BITBLT_MONO)
ge_cmd |= 0x00000002 | 0x00000100 | 0x00020000;
else
ge_cmd |= 0x00000001;
}
writel(ge_cmd, engine);
if (op == VIA_BITBLT_FILL || !src_mem)
return 0;
tmp = (width * height * (op == VIA_BITBLT_MONO ? 1 : (dst_bpp >> 3)) +
3) >> 2;
for (i = 0; i < tmp; i++)
writel(src_mem[i], engine + VIA_MMIO_BLTBASE);
return 0;
}
static int hw_bitblt_2(void __iomem *engine, u8 op, u32 width, u32 height,
u8 dst_bpp, u32 dst_addr, u32 dst_pitch, u32 dst_x, u32 dst_y,
u32 *src_mem, u32 src_addr, u32 src_pitch, u32 src_x, u32 src_y,
u32 fg_color, u32 bg_color, u8 fill_rop)
{
u32 ge_cmd = 0, tmp, i;
int ret;
if (!op || op > 3) {
printk(KERN_WARNING "hw_bitblt_2: Invalid operation: %d\n", op);
return -EINVAL;
}
if (op != VIA_BITBLT_FILL && !src_mem && src_addr == dst_addr) {
if (src_x < dst_x) {
ge_cmd |= 0x00008000;
src_x += width - 1;
dst_x += width - 1;
}
if (src_y < dst_y) {
ge_cmd |= 0x00004000;
src_y += height - 1;
dst_y += height - 1;
}
}
if (op == VIA_BITBLT_FILL) {
switch (fill_rop) {
case 0x00: /* blackness */
case 0x5A: /* pattern inversion */
case 0xF0: /* pattern copy */
case 0xFF: /* whiteness */
break;
default:
printk(KERN_WARNING "hw_bitblt_2: Invalid fill rop: "
"%u\n", fill_rop);
return -EINVAL;
}
}
ret = viafb_set_bpp(engine, dst_bpp);
if (ret)
return ret;
if (op == VIA_BITBLT_FILL)
tmp = 0;
else
tmp = src_pitch;
if (tmp & 0xFFFFC007 || dst_pitch & 0xFFFFC007) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported pitch %X %X\n",
tmp, dst_pitch);
return -EINVAL;
}
tmp = (tmp >> 3) | (dst_pitch << (16 - 3));
writel(tmp, engine + 0x08);
if ((width - 1) & 0xFFFFF000 || (height - 1) & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported width/height "
"%d %d\n", width, height);
return -EINVAL;
}
tmp = (width - 1) | ((height - 1) << 16);
writel(tmp, engine + 0x0C);
if (dst_x & 0xFFFFF000 || dst_y & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported destination x/y "
"%d %d\n", dst_x, dst_y);
return -EINVAL;
}
tmp = dst_x | (dst_y << 16);
writel(tmp, engine + 0x10);
if (dst_addr & 0xE0000007) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported destination "
"address %X\n", dst_addr);
return -EINVAL;
}
tmp = dst_addr >> 3;
writel(tmp, engine + 0x14);
if (op != VIA_BITBLT_FILL) {
if (src_x & (op == VIA_BITBLT_MONO ? 0xFFFF8000 : 0xFFFFF000)
|| src_y & 0xFFFFF000) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported source "
"x/y %d %d\n", src_x, src_y);
return -EINVAL;
}
tmp = src_x | (src_y << 16);
writel(tmp, engine + 0x18);
tmp = src_mem ? 0 : src_addr;
if (dst_addr & 0xE0000007) {
printk(KERN_WARNING "hw_bitblt_2: Unsupported source "
"address %X\n", tmp);
return -EINVAL;
}
tmp >>= 3;
writel(tmp, engine + 0x1C);
}
if (op == VIA_BITBLT_FILL) {
writel(fg_color, engine + 0x58);
} else if (op == VIA_BITBLT_MONO) {
writel(fg_color, engine + 0x4C);
writel(bg_color, engine + 0x50);
}
if (op == VIA_BITBLT_FILL)
ge_cmd |= fill_rop << 24 | 0x00002000 | 0x00000001;
else {
ge_cmd |= 0xCC000000; /* ROP=SRCCOPY */
if (src_mem)
ge_cmd |= 0x00000040;
if (op == VIA_BITBLT_MONO)
ge_cmd |= 0x00000002 | 0x00000100 | 0x00020000;
else
ge_cmd |= 0x00000001;
}
writel(ge_cmd, engine);
if (op == VIA_BITBLT_FILL || !src_mem)
return 0;
tmp = (width * height * (op == VIA_BITBLT_MONO ? 1 : (dst_bpp >> 3)) +
3) >> 2;
for (i = 0; i < tmp; i++)
writel(src_mem[i], engine + VIA_MMIO_BLTBASE);
return 0;
}
int viafb_setup_engine(struct fb_info *info)
{
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
struct viafb_par *viapar = info->par;
void __iomem *engine;
u32 chip_name = viapar->shared->chip_info.gfx_chip_name;
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
engine = viapar->shared->vdev->engine_mmio;
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
if (!engine) {
printk(KERN_WARNING "viafb_init_accel: ioremap failed, "
"hardware acceleration disabled\n");
return -ENOMEM;
}
switch (chip_name) {
case UNICHROME_CLE266:
case UNICHROME_K400:
case UNICHROME_K800:
case UNICHROME_PM800:
case UNICHROME_CN700:
case UNICHROME_CX700:
case UNICHROME_CN750:
case UNICHROME_K8M890:
case UNICHROME_P4M890:
case UNICHROME_P4M900:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
viapar->shared->hw_bitblt = hw_bitblt_1;
break;
case UNICHROME_VX800:
case UNICHROME_VX855:
case UNICHROME_VX900:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
viapar->shared->hw_bitblt = hw_bitblt_2;
break;
default:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
viapar->shared->hw_bitblt = NULL;
}
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
viapar->fbmem_free -= CURSOR_SIZE;
viapar->shared->cursor_vram_addr = viapar->fbmem_free;
viapar->fbmem_used += CURSOR_SIZE;
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
viapar->fbmem_free -= VQ_SIZE;
viapar->shared->vq_vram_addr = viapar->fbmem_free;
viapar->fbmem_used += VQ_SIZE;
#if defined(CONFIG_VIDEO_VIA_CAMERA) || defined(CONFIG_VIDEO_VIA_CAMERA_MODULE)
/*
* Set aside a chunk of framebuffer memory for the camera
* driver. Someday this driver probably needs a proper allocator
* for fbmem; for now, we just have to do this before the
* framebuffer initializes itself.
*
* As for the size: the engine can handle three frames,
* 16 bits deep, up to VGA resolution.
*/
viapar->shared->vdev->camera_fbmem_size = 3*VGA_HEIGHT*VGA_WIDTH*2;
viapar->fbmem_free -= viapar->shared->vdev->camera_fbmem_size;
viapar->fbmem_used += viapar->shared->vdev->camera_fbmem_size;
viapar->shared->vdev->camera_fbmem_offset = viapar->fbmem_free;
#endif
viafb_reset_engine(viapar);
return 0;
}
void viafb_reset_engine(struct viafb_par *viapar)
{
void __iomem *engine = viapar->shared->vdev->engine_mmio;
int highest_reg, i;
u32 vq_start_addr, vq_end_addr, vq_start_low, vq_end_low, vq_high,
vq_len, chip_name = viapar->shared->chip_info.gfx_chip_name;
/* Initialize registers to reset the 2D engine */
switch (viapar->shared->chip_info.twod_engine) {
case VIA_2D_ENG_M1:
highest_reg = 0x5c;
break;
default:
highest_reg = 0x40;
break;
}
for (i = 0; i <= highest_reg; i += 4)
writel(0x0, engine + i);
/* Init AGP and VQ regs */
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
switch (chip_name) {
case UNICHROME_K8M890:
case UNICHROME_P4M900:
case UNICHROME_VX800:
case UNICHROME_VX855:
case UNICHROME_VX900:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
writel(0x00100000, engine + VIA_REG_CR_TRANSET);
writel(0x680A0000, engine + VIA_REG_CR_TRANSPACE);
writel(0x02000000, engine + VIA_REG_CR_TRANSPACE);
break;
default:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
writel(0x00100000, engine + VIA_REG_TRANSET);
writel(0x00000000, engine + VIA_REG_TRANSPACE);
writel(0x00333004, engine + VIA_REG_TRANSPACE);
writel(0x60000000, engine + VIA_REG_TRANSPACE);
writel(0x61000000, engine + VIA_REG_TRANSPACE);
writel(0x62000000, engine + VIA_REG_TRANSPACE);
writel(0x63000000, engine + VIA_REG_TRANSPACE);
writel(0x64000000, engine + VIA_REG_TRANSPACE);
writel(0x7D000000, engine + VIA_REG_TRANSPACE);
writel(0xFE020000, engine + VIA_REG_TRANSET);
writel(0x00000000, engine + VIA_REG_TRANSPACE);
break;
}
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
/* Enable VQ */
vq_start_addr = viapar->shared->vq_vram_addr;
vq_end_addr = viapar->shared->vq_vram_addr + VQ_SIZE - 1;
vq_start_low = 0x50000000 | (vq_start_addr & 0xFFFFFF);
vq_end_low = 0x51000000 | (vq_end_addr & 0xFFFFFF);
vq_high = 0x52000000 | ((vq_start_addr & 0xFF000000) >> 24) |
((vq_end_addr & 0xFF000000) >> 16);
vq_len = 0x53000000 | (VQ_SIZE >> 3);
switch (chip_name) {
case UNICHROME_K8M890:
case UNICHROME_P4M900:
case UNICHROME_VX800:
case UNICHROME_VX855:
case UNICHROME_VX900:
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
vq_start_low |= 0x20000000;
vq_end_low |= 0x20000000;
vq_high |= 0x20000000;
vq_len |= 0x20000000;
writel(0x00100000, engine + VIA_REG_CR_TRANSET);
writel(vq_high, engine + VIA_REG_CR_TRANSPACE);
writel(vq_start_low, engine + VIA_REG_CR_TRANSPACE);
writel(vq_end_low, engine + VIA_REG_CR_TRANSPACE);
writel(vq_len, engine + VIA_REG_CR_TRANSPACE);
writel(0x74301001, engine + VIA_REG_CR_TRANSPACE);
writel(0x00000000, engine + VIA_REG_CR_TRANSPACE);
break;
default:
writel(0x00FE0000, engine + VIA_REG_TRANSET);
writel(0x080003FE, engine + VIA_REG_TRANSPACE);
writel(0x0A00027C, engine + VIA_REG_TRANSPACE);
writel(0x0B000260, engine + VIA_REG_TRANSPACE);
writel(0x0C000274, engine + VIA_REG_TRANSPACE);
writel(0x0D000264, engine + VIA_REG_TRANSPACE);
writel(0x0E000000, engine + VIA_REG_TRANSPACE);
writel(0x0F000020, engine + VIA_REG_TRANSPACE);
writel(0x1000027E, engine + VIA_REG_TRANSPACE);
writel(0x110002FE, engine + VIA_REG_TRANSPACE);
writel(0x200F0060, engine + VIA_REG_TRANSPACE);
writel(0x00000006, engine + VIA_REG_TRANSPACE);
writel(0x40008C0F, engine + VIA_REG_TRANSPACE);
writel(0x44000000, engine + VIA_REG_TRANSPACE);
writel(0x45080C04, engine + VIA_REG_TRANSPACE);
writel(0x46800408, engine + VIA_REG_TRANSPACE);
writel(vq_high, engine + VIA_REG_TRANSPACE);
writel(vq_start_low, engine + VIA_REG_TRANSPACE);
writel(vq_end_low, engine + VIA_REG_TRANSPACE);
writel(vq_len, engine + VIA_REG_TRANSPACE);
break;
}
/* Set Cursor Image Base Address */
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
writel(viapar->shared->cursor_vram_addr, engine + VIA_REG_CURSOR_MODE);
writel(0x0, engine + VIA_REG_CURSOR_POS);
writel(0x0, engine + VIA_REG_CURSOR_ORG);
writel(0x0, engine + VIA_REG_CURSOR_BG);
writel(0x0, engine + VIA_REG_CURSOR_FG);
return;
}
void viafb_show_hw_cursor(struct fb_info *info, int Status)
{
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
struct viafb_par *viapar = info->par;
u32 temp, iga_path = viapar->iga_path;
temp = readl(viapar->shared->vdev->engine_mmio + VIA_REG_CURSOR_MODE);
switch (Status) {
case HW_Cursor_ON:
temp |= 0x1;
break;
case HW_Cursor_OFF:
temp &= 0xFFFFFFFE;
break;
}
switch (iga_path) {
case IGA2:
temp |= 0x80000000;
break;
case IGA1:
default:
temp &= 0x7FFFFFFF;
}
writel(temp, viapar->shared->vdev->engine_mmio + VIA_REG_CURSOR_MODE);
}
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
void viafb_wait_engine_idle(struct fb_info *info)
{
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
struct viafb_par *viapar = info->par;
int loop = 0;
u32 mask;
void __iomem *engine = viapar->shared->vdev->engine_mmio;
switch (viapar->shared->chip_info.twod_engine) {
case VIA_2D_ENG_H5:
case VIA_2D_ENG_M1:
mask = VIA_CMD_RGTR_BUSY_M1 | VIA_2D_ENG_BUSY_M1 |
VIA_3D_ENG_BUSY_M1;
break;
default:
while (!(readl(engine + VIA_REG_STATUS) &
VIA_VR_QUEUE_BUSY) && (loop < MAXLOOP)) {
loop++;
cpu_relax();
}
mask = VIA_CMD_RGTR_BUSY | VIA_2D_ENG_BUSY | VIA_3D_ENG_BUSY;
break;
}
while ((readl(engine + VIA_REG_STATUS) & mask) && (loop < MAXLOOP)) {
loop++;
cpu_relax();
}
viafb: hardware acceleration initialization cleanup The main motivation of this patch was to merge the three initialization functions in one and clean it up. However as some changes in other code areas where needed to do it right some small other changes were made. Changes to viafb_par: io_virt renamed as engine_mmio and moved to shared VQ_start renamed as vq_vram_addr and moved to shared VQ_end removed as it is easily recalculatable vq_vram_addr is not strictly needed but keep it to track where we allocated video memory. The memory allocated for the virtual queue was shrunk to VQ_SIZE as VQ_SIZE+CURSOR_SIZE looked like a bug to me. But to be honest I don't have the faintest idea what virtual queues are for in the graphic hardware and whether the driver needs them in any way. I only know that they aren't directly accessed by the driver and so the only potential current use would be as hardware internal buffers. For now keep them to avoid regressions and only remove the double cursor allocation. The most changes were caused by renames and the mentioned structure changes so the chance of regressions is pretty low. The meaning of viafb_accel changed slightly as previously it was changed back and forth in the code and allowed to enable the hardware acceleration by software if previously disabled. The new behaviour is that viafb_accel=0 always prevents hardware acceleration. With viafb_accel!=0 the acceleration can be freely choosen by set_var. This means viafb_accel is a diagnostic tool and if someone has to use viafb_accel=0 the driver needs to be fixed. As this is mostly a code cleanup no regressions beside the slightly change of viafb_accel is expected. Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Scott Fang <ScottFang@viatech.com.cn> Cc: Joseph Chan <JosephChan@via.com.tw> Cc: Harald Welte <laforge@gnumonks.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-09-23 07:47:31 +08:00
if (loop >= MAXLOOP)
printk(KERN_ERR "viafb_wait_engine_idle: not syncing\n");
}