linux-sg2042/drivers/video/pxafb.c

2297 lines
60 KiB
C
Raw Normal View History

/*
* linux/drivers/video/pxafb.c
*
* Copyright (C) 1999 Eric A. Thomas.
* Copyright (C) 2004 Jean-Frederic Clere.
* Copyright (C) 2004 Ian Campbell.
* Copyright (C) 2004 Jeff Lackey.
* Based on sa1100fb.c Copyright (C) 1999 Eric A. Thomas
* which in turn is
* Based on acornfb.c Copyright (C) Russell King.
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file COPYING in the main directory of this archive for
* more details.
*
* Intel PXA250/210 LCD Controller Frame Buffer Driver
*
* Please direct your questions and comments on this driver to the following
* email address:
*
* linux-arm-kernel@lists.arm.linux.org.uk
*
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
* Add support for overlay1 and overlay2 based on pxafb_overlay.c:
*
* Copyright (C) 2004, Intel Corporation
*
* 2003/08/27: <yu.tang@intel.com>
* 2004/03/10: <stanley.cai@intel.com>
* 2004/10/28: <yan.yin@intel.com>
*
* Copyright (C) 2006-2008 Marvell International Ltd.
* All Rights Reserved
*/
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/interrupt.h>
#include <linux/slab.h>
2008-07-24 12:28:13 +08:00
#include <linux/mm.h>
#include <linux/fb.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/ioport.h>
#include <linux/cpufreq.h>
#include <linux/platform_device.h>
#include <linux/dma-mapping.h>
#include <linux/clk.h>
#include <linux/err.h>
#include <linux/completion.h>
#include <linux/mutex.h>
#include <linux/kthread.h>
#include <linux/freezer.h>
#include <mach/hardware.h>
#include <asm/io.h>
#include <asm/irq.h>
#include <asm/div64.h>
#include <mach/bitfield.h>
#include <mach/pxafb.h>
/*
* Complain if VAR is out of range.
*/
#define DEBUG_VAR 1
#include "pxafb.h"
/* Bits which should not be set in machine configuration structures */
#define LCCR0_INVALID_CONFIG_MASK (LCCR0_OUM | LCCR0_BM | LCCR0_QDM |\
LCCR0_DIS | LCCR0_EFM | LCCR0_IUM |\
LCCR0_SFM | LCCR0_LDM | LCCR0_ENB)
#define LCCR3_INVALID_CONFIG_MASK (LCCR3_HSP | LCCR3_VSP |\
LCCR3_PCD | LCCR3_BPP(0xf))
static int pxafb_activate_var(struct fb_var_screeninfo *var,
struct pxafb_info *);
static void set_ctrlr_state(struct pxafb_info *fbi, u_int state);
static void setup_base_frame(struct pxafb_info *fbi,
struct fb_var_screeninfo *var, int branch);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
static int setup_frame_dma(struct pxafb_info *fbi, int dma, int pal,
unsigned long offset, size_t size);
static unsigned long video_mem_size = 0;
static inline unsigned long
lcd_readl(struct pxafb_info *fbi, unsigned int off)
{
return __raw_readl(fbi->mmio_base + off);
}
static inline void
lcd_writel(struct pxafb_info *fbi, unsigned int off, unsigned long val)
{
__raw_writel(val, fbi->mmio_base + off);
}
static inline void pxafb_schedule_work(struct pxafb_info *fbi, u_int state)
{
unsigned long flags;
local_irq_save(flags);
/*
* We need to handle two requests being made at the same time.
* There are two important cases:
* 1. When we are changing VT (C_REENABLE) while unblanking
* (C_ENABLE) We must perform the unblanking, which will
* do our REENABLE for us.
* 2. When we are blanking, but immediately unblank before
* we have blanked. We do the "REENABLE" thing here as
* well, just to be sure.
*/
if (fbi->task_state == C_ENABLE && state == C_REENABLE)
state = (u_int) -1;
if (fbi->task_state == C_DISABLE && state == C_ENABLE)
state = C_REENABLE;
if (state != (u_int)-1) {
fbi->task_state = state;
schedule_work(&fbi->task);
}
local_irq_restore(flags);
}
static inline u_int chan_to_field(u_int chan, struct fb_bitfield *bf)
{
chan &= 0xffff;
chan >>= 16 - bf->length;
return chan << bf->offset;
}
static int
pxafb_setpalettereg(u_int regno, u_int red, u_int green, u_int blue,
u_int trans, struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
u_int val;
if (regno >= fbi->palette_size)
return 1;
if (fbi->fb.var.grayscale) {
fbi->palette_cpu[regno] = ((blue >> 8) & 0x00ff);
return 0;
}
switch (fbi->lccr4 & LCCR4_PAL_FOR_MASK) {
case LCCR4_PAL_FOR_0:
val = ((red >> 0) & 0xf800);
val |= ((green >> 5) & 0x07e0);
val |= ((blue >> 11) & 0x001f);
fbi->palette_cpu[regno] = val;
break;
case LCCR4_PAL_FOR_1:
val = ((red << 8) & 0x00f80000);
val |= ((green >> 0) & 0x0000fc00);
val |= ((blue >> 8) & 0x000000f8);
((u32 *)(fbi->palette_cpu))[regno] = val;
break;
case LCCR4_PAL_FOR_2:
val = ((red << 8) & 0x00fc0000);
val |= ((green >> 0) & 0x0000fc00);
val |= ((blue >> 8) & 0x000000fc);
((u32 *)(fbi->palette_cpu))[regno] = val;
break;
case LCCR4_PAL_FOR_3:
val = ((red << 8) & 0x00ff0000);
val |= ((green >> 0) & 0x0000ff00);
val |= ((blue >> 8) & 0x000000ff);
((u32 *)(fbi->palette_cpu))[regno] = val;
break;
}
return 0;
}
static int
pxafb_setcolreg(u_int regno, u_int red, u_int green, u_int blue,
u_int trans, struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
unsigned int val;
int ret = 1;
/*
* If inverse mode was selected, invert all the colours
* rather than the register number. The register number
* is what you poke into the framebuffer to produce the
* colour you requested.
*/
if (fbi->cmap_inverse) {
red = 0xffff - red;
green = 0xffff - green;
blue = 0xffff - blue;
}
/*
* If greyscale is true, then we convert the RGB value
* to greyscale no matter what visual we are using.
*/
if (fbi->fb.var.grayscale)
red = green = blue = (19595 * red + 38470 * green +
7471 * blue) >> 16;
switch (fbi->fb.fix.visual) {
case FB_VISUAL_TRUECOLOR:
/*
* 16-bit True Colour. We encode the RGB value
* according to the RGB bitfield information.
*/
if (regno < 16) {
u32 *pal = fbi->fb.pseudo_palette;
val = chan_to_field(red, &fbi->fb.var.red);
val |= chan_to_field(green, &fbi->fb.var.green);
val |= chan_to_field(blue, &fbi->fb.var.blue);
pal[regno] = val;
ret = 0;
}
break;
case FB_VISUAL_STATIC_PSEUDOCOLOR:
case FB_VISUAL_PSEUDOCOLOR:
ret = pxafb_setpalettereg(regno, red, green, blue, trans, info);
break;
}
return ret;
}
/* calculate pixel depth, transparency bit included, >=16bpp formats _only_ */
static inline int var_to_depth(struct fb_var_screeninfo *var)
{
return var->red.length + var->green.length +
var->blue.length + var->transp.length;
}
/* calculate 4-bit BPP value for LCCR3 and OVLxC1 */
static int pxafb_var_to_bpp(struct fb_var_screeninfo *var)
{
int bpp = -EINVAL;
switch (var->bits_per_pixel) {
case 1: bpp = 0; break;
case 2: bpp = 1; break;
case 4: bpp = 2; break;
case 8: bpp = 3; break;
case 16: bpp = 4; break;
case 24:
switch (var_to_depth(var)) {
case 18: bpp = 6; break; /* 18-bits/pixel packed */
case 19: bpp = 8; break; /* 19-bits/pixel packed */
case 24: bpp = 9; break;
}
break;
case 32:
switch (var_to_depth(var)) {
case 18: bpp = 5; break; /* 18-bits/pixel unpacked */
case 19: bpp = 7; break; /* 19-bits/pixel unpacked */
case 25: bpp = 10; break;
}
break;
}
return bpp;
}
/*
* pxafb_var_to_lccr3():
* Convert a bits per pixel value to the correct bit pattern for LCCR3
*
* NOTE: for PXA27x with overlays support, the LCCR3_PDFOR_x bits have an
* implication of the acutal use of transparency bit, which we handle it
* here separatedly. See PXA27x Developer's Manual, Section <<7.4.6 Pixel
* Formats>> for the valid combination of PDFOR, PAL_FOR for various BPP.
*
* Transparency for palette pixel formats is not supported at the moment.
*/
static uint32_t pxafb_var_to_lccr3(struct fb_var_screeninfo *var)
{
int bpp = pxafb_var_to_bpp(var);
uint32_t lccr3;
if (bpp < 0)
return 0;
lccr3 = LCCR3_BPP(bpp);
switch (var_to_depth(var)) {
case 16: lccr3 |= var->transp.length ? LCCR3_PDFOR_3 : 0; break;
case 18: lccr3 |= LCCR3_PDFOR_3; break;
case 24: lccr3 |= var->transp.length ? LCCR3_PDFOR_2 : LCCR3_PDFOR_3;
break;
case 19:
case 25: lccr3 |= LCCR3_PDFOR_0; break;
}
return lccr3;
}
#define SET_PIXFMT(v, r, g, b, t) \
({ \
(v)->transp.offset = (t) ? (r) + (g) + (b) : 0; \
(v)->transp.length = (t) ? (t) : 0; \
(v)->blue.length = (b); (v)->blue.offset = 0; \
(v)->green.length = (g); (v)->green.offset = (b); \
(v)->red.length = (r); (v)->red.offset = (b) + (g); \
})
/* set the RGBT bitfields of fb_var_screeninf according to
* var->bits_per_pixel and given depth
*/
static void pxafb_set_pixfmt(struct fb_var_screeninfo *var, int depth)
{
if (depth == 0)
depth = var->bits_per_pixel;
if (var->bits_per_pixel < 16) {
/* indexed pixel formats */
var->red.offset = 0; var->red.length = 8;
var->green.offset = 0; var->green.length = 8;
var->blue.offset = 0; var->blue.length = 8;
var->transp.offset = 0; var->transp.length = 8;
}
switch (depth) {
case 16: var->transp.length ?
SET_PIXFMT(var, 5, 5, 5, 1) : /* RGBT555 */
SET_PIXFMT(var, 5, 6, 5, 0); break; /* RGB565 */
case 18: SET_PIXFMT(var, 6, 6, 6, 0); break; /* RGB666 */
case 19: SET_PIXFMT(var, 6, 6, 6, 1); break; /* RGBT666 */
case 24: var->transp.length ?
SET_PIXFMT(var, 8, 8, 7, 1) : /* RGBT887 */
SET_PIXFMT(var, 8, 8, 8, 0); break; /* RGB888 */
case 25: SET_PIXFMT(var, 8, 8, 8, 1); break; /* RGBT888 */
}
}
#ifdef CONFIG_CPU_FREQ
/*
* pxafb_display_dma_period()
* Calculate the minimum period (in picoseconds) between two DMA
* requests for the LCD controller. If we hit this, it means we're
* doing nothing but LCD DMA.
*/
static unsigned int pxafb_display_dma_period(struct fb_var_screeninfo *var)
{
/*
* Period = pixclock * bits_per_byte * bytes_per_transfer
* / memory_bits_per_pixel;
*/
return var->pixclock * 8 * 16 / var->bits_per_pixel;
}
#endif
/*
* Select the smallest mode that allows the desired resolution to be
* displayed. If desired parameters can be rounded up.
*/
static struct pxafb_mode_info *pxafb_getmode(struct pxafb_mach_info *mach,
struct fb_var_screeninfo *var)
{
struct pxafb_mode_info *mode = NULL;
struct pxafb_mode_info *modelist = mach->modes;
unsigned int best_x = 0xffffffff, best_y = 0xffffffff;
unsigned int i;
for (i = 0; i < mach->num_modes; i++) {
if (modelist[i].xres >= var->xres &&
modelist[i].yres >= var->yres &&
modelist[i].xres < best_x &&
modelist[i].yres < best_y &&
modelist[i].bpp >= var->bits_per_pixel) {
best_x = modelist[i].xres;
best_y = modelist[i].yres;
mode = &modelist[i];
}
}
return mode;
}
static void pxafb_setmode(struct fb_var_screeninfo *var,
struct pxafb_mode_info *mode)
{
var->xres = mode->xres;
var->yres = mode->yres;
var->bits_per_pixel = mode->bpp;
var->pixclock = mode->pixclock;
var->hsync_len = mode->hsync_len;
var->left_margin = mode->left_margin;
var->right_margin = mode->right_margin;
var->vsync_len = mode->vsync_len;
var->upper_margin = mode->upper_margin;
var->lower_margin = mode->lower_margin;
var->sync = mode->sync;
var->grayscale = mode->cmap_greyscale;
var->transp.length = mode->transparency;
/* set the initial RGBA bitfields */
pxafb_set_pixfmt(var, mode->depth);
}
static int pxafb_adjust_timing(struct pxafb_info *fbi,
struct fb_var_screeninfo *var)
{
int line_length;
var->xres = max_t(int, var->xres, MIN_XRES);
var->yres = max_t(int, var->yres, MIN_YRES);
if (!(fbi->lccr0 & LCCR0_LCDT)) {
clamp_val(var->hsync_len, 1, 64);
clamp_val(var->vsync_len, 1, 64);
clamp_val(var->left_margin, 1, 255);
clamp_val(var->right_margin, 1, 255);
clamp_val(var->upper_margin, 1, 255);
clamp_val(var->lower_margin, 1, 255);
}
/* make sure each line is aligned on word boundary */
line_length = var->xres * var->bits_per_pixel / 8;
line_length = ALIGN(line_length, 4);
var->xres = line_length * 8 / var->bits_per_pixel;
/* we don't support xpan, force xres_virtual to be equal to xres */
var->xres_virtual = var->xres;
if (var->accel_flags & FB_ACCELF_TEXT)
var->yres_virtual = fbi->fb.fix.smem_len / line_length;
else
var->yres_virtual = max(var->yres_virtual, var->yres);
/* check for limits */
if (var->xres > MAX_XRES || var->yres > MAX_YRES)
return -EINVAL;
if (var->yres > var->yres_virtual)
return -EINVAL;
return 0;
}
/*
* pxafb_check_var():
* Get the video params out of 'var'. If a value doesn't fit, round it up,
* if it's too big, return -EINVAL.
*
* Round up in the following order: bits_per_pixel, xres,
* yres, xres_virtual, yres_virtual, xoffset, yoffset, grayscale,
* bitfields, horizontal timing, vertical timing.
*/
static int pxafb_check_var(struct fb_var_screeninfo *var, struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
struct pxafb_mach_info *inf = fbi->dev->platform_data;
int err;
if (inf->fixed_modes) {
struct pxafb_mode_info *mode;
mode = pxafb_getmode(inf, var);
if (!mode)
return -EINVAL;
pxafb_setmode(var, mode);
}
/* do a test conversion to BPP fields to check the color formats */
err = pxafb_var_to_bpp(var);
if (err < 0)
return err;
pxafb_set_pixfmt(var, var_to_depth(var));
err = pxafb_adjust_timing(fbi, var);
if (err)
return err;
#ifdef CONFIG_CPU_FREQ
pr_debug("pxafb: dma period = %d ps\n",
pxafb_display_dma_period(var));
#endif
return 0;
}
/*
* pxafb_set_par():
* Set the user defined part of the display for the specified console
*/
static int pxafb_set_par(struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
struct fb_var_screeninfo *var = &info->var;
if (var->bits_per_pixel >= 16)
fbi->fb.fix.visual = FB_VISUAL_TRUECOLOR;
else if (!fbi->cmap_static)
fbi->fb.fix.visual = FB_VISUAL_PSEUDOCOLOR;
else {
/*
* Some people have weird ideas about wanting static
* pseudocolor maps. I suspect their user space
* applications are broken.
*/
fbi->fb.fix.visual = FB_VISUAL_STATIC_PSEUDOCOLOR;
}
fbi->fb.fix.line_length = var->xres_virtual *
var->bits_per_pixel / 8;
if (var->bits_per_pixel >= 16)
fbi->palette_size = 0;
else
fbi->palette_size = var->bits_per_pixel == 1 ?
4 : 1 << var->bits_per_pixel;
fbi->palette_cpu = (u16 *)&fbi->dma_buff->palette[0];
if (fbi->fb.var.bits_per_pixel >= 16)
fb_dealloc_cmap(&fbi->fb.cmap);
else
fb_alloc_cmap(&fbi->fb.cmap, 1<<fbi->fb.var.bits_per_pixel, 0);
pxafb_activate_var(var, fbi);
return 0;
}
static int pxafb_pan_display(struct fb_var_screeninfo *var,
struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
struct fb_var_screeninfo newvar;
int dma = DMA_MAX + DMA_BASE;
if (fbi->state != C_ENABLE)
return 0;
/* Only take .xoffset, .yoffset and .vmode & FB_VMODE_YWRAP from what
* was passed in and copy the rest from the old screeninfo.
*/
memcpy(&newvar, &fbi->fb.var, sizeof(newvar));
newvar.xoffset = var->xoffset;
newvar.yoffset = var->yoffset;
newvar.vmode &= ~FB_VMODE_YWRAP;
newvar.vmode |= var->vmode & FB_VMODE_YWRAP;
setup_base_frame(fbi, &newvar, 1);
if (fbi->lccr0 & LCCR0_SDS)
lcd_writel(fbi, FBR1, fbi->fdadr[dma + 1] | 0x1);
lcd_writel(fbi, FBR0, fbi->fdadr[dma] | 0x1);
return 0;
}
/*
* pxafb_blank():
* Blank the display by setting all palette values to zero. Note, the
* 16 bpp mode does not really use the palette, so this will not
* blank the display in all modes.
*/
static int pxafb_blank(int blank, struct fb_info *info)
{
struct pxafb_info *fbi = (struct pxafb_info *)info;
int i;
switch (blank) {
case FB_BLANK_POWERDOWN:
case FB_BLANK_VSYNC_SUSPEND:
case FB_BLANK_HSYNC_SUSPEND:
case FB_BLANK_NORMAL:
if (fbi->fb.fix.visual == FB_VISUAL_PSEUDOCOLOR ||
fbi->fb.fix.visual == FB_VISUAL_STATIC_PSEUDOCOLOR)
for (i = 0; i < fbi->palette_size; i++)
pxafb_setpalettereg(i, 0, 0, 0, 0, info);
pxafb_schedule_work(fbi, C_DISABLE);
/* TODO if (pxafb_blank_helper) pxafb_blank_helper(blank); */
break;
case FB_BLANK_UNBLANK:
/* TODO if (pxafb_blank_helper) pxafb_blank_helper(blank); */
if (fbi->fb.fix.visual == FB_VISUAL_PSEUDOCOLOR ||
fbi->fb.fix.visual == FB_VISUAL_STATIC_PSEUDOCOLOR)
fb_set_cmap(&fbi->fb.cmap, info);
pxafb_schedule_work(fbi, C_ENABLE);
}
return 0;
}
static struct fb_ops pxafb_ops = {
.owner = THIS_MODULE,
.fb_check_var = pxafb_check_var,
.fb_set_par = pxafb_set_par,
.fb_pan_display = pxafb_pan_display,
.fb_setcolreg = pxafb_setcolreg,
.fb_fillrect = cfb_fillrect,
.fb_copyarea = cfb_copyarea,
.fb_imageblit = cfb_imageblit,
.fb_blank = pxafb_blank,
};
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
#ifdef CONFIG_FB_PXA_OVERLAY
static void overlay1fb_setup(struct pxafb_layer *ofb)
{
int size = ofb->fb.fix.line_length * ofb->fb.var.yres_virtual;
unsigned long start = ofb->video_mem_phys;
setup_frame_dma(ofb->fbi, DMA_OV1, PAL_NONE, start, size);
}
/* Depending on the enable status of overlay1/2, the DMA should be
* updated from FDADRx (when disabled) or FBRx (when enabled).
*/
static void overlay1fb_enable(struct pxafb_layer *ofb)
{
int enabled = lcd_readl(ofb->fbi, OVL1C1) & OVLxC1_OEN;
uint32_t fdadr1 = ofb->fbi->fdadr[DMA_OV1] | (enabled ? 0x1 : 0);
lcd_writel(ofb->fbi, enabled ? FBR1 : FDADR1, fdadr1);
lcd_writel(ofb->fbi, OVL1C2, ofb->control[1]);
lcd_writel(ofb->fbi, OVL1C1, ofb->control[0] | OVLxC1_OEN);
}
static void overlay1fb_disable(struct pxafb_layer *ofb)
{
uint32_t lccr5 = lcd_readl(ofb->fbi, LCCR5);
lcd_writel(ofb->fbi, OVL1C1, ofb->control[0] & ~OVLxC1_OEN);
lcd_writel(ofb->fbi, LCSR1, LCSR1_BS(1));
lcd_writel(ofb->fbi, LCCR5, lccr5 & ~LCSR1_BS(1));
lcd_writel(ofb->fbi, FBR1, ofb->fbi->fdadr[DMA_OV1] | 0x3);
if (wait_for_completion_timeout(&ofb->branch_done, 1 * HZ) == 0)
pr_warning("%s: timeout disabling overlay1\n", __func__);
lcd_writel(ofb->fbi, LCCR5, lccr5);
}
static void overlay2fb_setup(struct pxafb_layer *ofb)
{
int size, div = 1, pfor = NONSTD_TO_PFOR(ofb->fb.var.nonstd);
unsigned long start[3] = { ofb->video_mem_phys, 0, 0 };
if (pfor == OVERLAY_FORMAT_RGB || pfor == OVERLAY_FORMAT_YUV444_PACKED) {
size = ofb->fb.fix.line_length * ofb->fb.var.yres_virtual;
setup_frame_dma(ofb->fbi, DMA_OV2_Y, -1, start[0], size);
} else {
size = ofb->fb.var.xres_virtual * ofb->fb.var.yres_virtual;
switch (pfor) {
case OVERLAY_FORMAT_YUV444_PLANAR: div = 1; break;
case OVERLAY_FORMAT_YUV422_PLANAR: div = 2; break;
case OVERLAY_FORMAT_YUV420_PLANAR: div = 4; break;
}
start[1] = start[0] + size;
start[2] = start[1] + size / div;
setup_frame_dma(ofb->fbi, DMA_OV2_Y, -1, start[0], size);
setup_frame_dma(ofb->fbi, DMA_OV2_Cb, -1, start[1], size / div);
setup_frame_dma(ofb->fbi, DMA_OV2_Cr, -1, start[2], size / div);
}
}
static void overlay2fb_enable(struct pxafb_layer *ofb)
{
int pfor = NONSTD_TO_PFOR(ofb->fb.var.nonstd);
int enabled = lcd_readl(ofb->fbi, OVL2C1) & OVLxC1_OEN;
uint32_t fdadr2 = ofb->fbi->fdadr[DMA_OV2_Y] | (enabled ? 0x1 : 0);
uint32_t fdadr3 = ofb->fbi->fdadr[DMA_OV2_Cb] | (enabled ? 0x1 : 0);
uint32_t fdadr4 = ofb->fbi->fdadr[DMA_OV2_Cr] | (enabled ? 0x1 : 0);
if (pfor == OVERLAY_FORMAT_RGB || pfor == OVERLAY_FORMAT_YUV444_PACKED)
lcd_writel(ofb->fbi, enabled ? FBR2 : FDADR2, fdadr2);
else {
lcd_writel(ofb->fbi, enabled ? FBR2 : FDADR2, fdadr2);
lcd_writel(ofb->fbi, enabled ? FBR3 : FDADR3, fdadr3);
lcd_writel(ofb->fbi, enabled ? FBR4 : FDADR4, fdadr4);
}
lcd_writel(ofb->fbi, OVL2C2, ofb->control[1]);
lcd_writel(ofb->fbi, OVL2C1, ofb->control[0] | OVLxC1_OEN);
}
static void overlay2fb_disable(struct pxafb_layer *ofb)
{
uint32_t lccr5 = lcd_readl(ofb->fbi, LCCR5);
lcd_writel(ofb->fbi, OVL2C1, ofb->control[0] & ~OVLxC1_OEN);
lcd_writel(ofb->fbi, LCSR1, LCSR1_BS(2));
lcd_writel(ofb->fbi, LCCR5, lccr5 & ~LCSR1_BS(2));
lcd_writel(ofb->fbi, FBR2, ofb->fbi->fdadr[DMA_OV2_Y] | 0x3);
lcd_writel(ofb->fbi, FBR3, ofb->fbi->fdadr[DMA_OV2_Cb] | 0x3);
lcd_writel(ofb->fbi, FBR4, ofb->fbi->fdadr[DMA_OV2_Cr] | 0x3);
if (wait_for_completion_timeout(&ofb->branch_done, 1 * HZ) == 0)
pr_warning("%s: timeout disabling overlay2\n", __func__);
}
static struct pxafb_layer_ops ofb_ops[] = {
[0] = {
.enable = overlay1fb_enable,
.disable = overlay1fb_disable,
.setup = overlay1fb_setup,
},
[1] = {
.enable = overlay2fb_enable,
.disable = overlay2fb_disable,
.setup = overlay2fb_setup,
},
};
static int overlayfb_open(struct fb_info *info, int user)
{
struct pxafb_layer *ofb = (struct pxafb_layer *)info;
/* no support for framebuffer console on overlay */
if (user == 0)
return -ENODEV;
/* allow only one user at a time */
if (atomic_inc_and_test(&ofb->usage))
return -EBUSY;
/* unblank the base framebuffer */
fb_blank(&ofb->fbi->fb, FB_BLANK_UNBLANK);
return 0;
}
static int overlayfb_release(struct fb_info *info, int user)
{
struct pxafb_layer *ofb = (struct pxafb_layer*) info;
atomic_dec(&ofb->usage);
ofb->ops->disable(ofb);
free_pages_exact(ofb->video_mem, ofb->video_mem_size);
ofb->video_mem = NULL;
ofb->video_mem_size = 0;
return 0;
}
static int overlayfb_check_var(struct fb_var_screeninfo *var,
struct fb_info *info)
{
struct pxafb_layer *ofb = (struct pxafb_layer *)info;
struct fb_var_screeninfo *base_var = &ofb->fbi->fb.var;
int xpos, ypos, pfor, bpp;
xpos = NONSTD_TO_XPOS(var->nonstd);
ypos = NONSTD_TO_XPOS(var->nonstd);
pfor = NONSTD_TO_PFOR(var->nonstd);
bpp = pxafb_var_to_bpp(var);
if (bpp < 0)
return -EINVAL;
/* no support for YUV format on overlay1 */
if (ofb->id == OVERLAY1 && pfor != 0)
return -EINVAL;
/* for YUV packed formats, bpp = 'minimum bpp of YUV components' */
switch (pfor) {
case OVERLAY_FORMAT_RGB:
bpp = pxafb_var_to_bpp(var);
if (bpp < 0)
return -EINVAL;
pxafb_set_pixfmt(var, var_to_depth(var));
break;
case OVERLAY_FORMAT_YUV444_PACKED: bpp = 24; break;
case OVERLAY_FORMAT_YUV444_PLANAR: bpp = 8; break;
case OVERLAY_FORMAT_YUV422_PLANAR: bpp = 4; break;
case OVERLAY_FORMAT_YUV420_PLANAR: bpp = 2; break;
default:
return -EINVAL;
}
/* each line must start at a 32-bit word boundary */
if ((xpos * bpp) % 32)
return -EINVAL;
/* xres must align on 32-bit word boundary */
var->xres = roundup(var->xres * bpp, 32) / bpp;
if ((xpos + var->xres > base_var->xres) ||
(ypos + var->yres > base_var->yres))
return -EINVAL;
var->xres_virtual = var->xres;
var->yres_virtual = max(var->yres, var->yres_virtual);
return 0;
}
static int overlayfb_map_video_memory(struct pxafb_layer *ofb)
{
struct fb_var_screeninfo *var = &ofb->fb.var;
int pfor = NONSTD_TO_PFOR(var->nonstd);
int size, bpp = 0;
switch (pfor) {
case OVERLAY_FORMAT_RGB: bpp = var->bits_per_pixel; break;
case OVERLAY_FORMAT_YUV444_PACKED: bpp = 24; break;
case OVERLAY_FORMAT_YUV444_PLANAR: bpp = 24; break;
case OVERLAY_FORMAT_YUV422_PLANAR: bpp = 16; break;
case OVERLAY_FORMAT_YUV420_PLANAR: bpp = 12; break;
}
ofb->fb.fix.line_length = var->xres_virtual * bpp / 8;
size = PAGE_ALIGN(ofb->fb.fix.line_length * var->yres_virtual);
/* don't re-allocate if the original video memory is enough */
if (ofb->video_mem) {
if (ofb->video_mem_size >= size)
return 0;
free_pages_exact(ofb->video_mem, ofb->video_mem_size);
}
ofb->video_mem = alloc_pages_exact(size, GFP_KERNEL | __GFP_ZERO);
if (ofb->video_mem == NULL)
return -ENOMEM;
ofb->video_mem_phys = virt_to_phys(ofb->video_mem);
ofb->video_mem_size = size;
mutex_lock(&ofb->fb.mm_lock);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
ofb->fb.fix.smem_start = ofb->video_mem_phys;
ofb->fb.fix.smem_len = ofb->fb.fix.line_length * var->yres_virtual;
mutex_unlock(&ofb->fb.mm_lock);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
ofb->fb.screen_base = ofb->video_mem;
return 0;
}
static int overlayfb_set_par(struct fb_info *info)
{
struct pxafb_layer *ofb = (struct pxafb_layer *)info;
struct fb_var_screeninfo *var = &info->var;
int xpos, ypos, pfor, bpp, ret;
ret = overlayfb_map_video_memory(ofb);
if (ret)
return ret;
bpp = pxafb_var_to_bpp(var);
xpos = NONSTD_TO_XPOS(var->nonstd);
ypos = NONSTD_TO_XPOS(var->nonstd);
pfor = NONSTD_TO_PFOR(var->nonstd);
ofb->control[0] = OVLxC1_PPL(var->xres) | OVLxC1_LPO(var->yres) |
OVLxC1_BPP(bpp);
ofb->control[1] = OVLxC2_XPOS(xpos) | OVLxC2_YPOS(ypos);
if (ofb->id == OVERLAY2)
ofb->control[1] |= OVL2C2_PFOR(pfor);
ofb->ops->setup(ofb);
ofb->ops->enable(ofb);
return 0;
}
static struct fb_ops overlay_fb_ops = {
.owner = THIS_MODULE,
.fb_open = overlayfb_open,
.fb_release = overlayfb_release,
.fb_check_var = overlayfb_check_var,
.fb_set_par = overlayfb_set_par,
};
static void __devinit init_pxafb_overlay(struct pxafb_info *fbi,
struct pxafb_layer *ofb, int id)
{
sprintf(ofb->fb.fix.id, "overlay%d", id + 1);
ofb->fb.fix.type = FB_TYPE_PACKED_PIXELS;
ofb->fb.fix.xpanstep = 0;
ofb->fb.fix.ypanstep = 1;
ofb->fb.var.activate = FB_ACTIVATE_NOW;
ofb->fb.var.height = -1;
ofb->fb.var.width = -1;
ofb->fb.var.vmode = FB_VMODE_NONINTERLACED;
ofb->fb.fbops = &overlay_fb_ops;
ofb->fb.flags = FBINFO_FLAG_DEFAULT;
ofb->fb.node = -1;
ofb->fb.pseudo_palette = NULL;
ofb->id = id;
ofb->ops = &ofb_ops[id];
atomic_set(&ofb->usage, 0);
ofb->fbi = fbi;
init_completion(&ofb->branch_done);
}
static inline int pxafb_overlay_supported(void)
{
if (cpu_is_pxa27x() || cpu_is_pxa3xx())
return 1;
return 0;
}
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
static int __devinit pxafb_overlay_init(struct pxafb_info *fbi)
{
int i, ret;
if (!pxafb_overlay_supported())
return 0;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
for (i = 0; i < 2; i++) {
init_pxafb_overlay(fbi, &fbi->overlay[i], i);
ret = register_framebuffer(&fbi->overlay[i].fb);
if (ret) {
dev_err(fbi->dev, "failed to register overlay %d\n", i);
return ret;
}
}
/* mask all IU/BS/EOF/SOF interrupts */
lcd_writel(fbi, LCCR5, ~0);
/* place overlay(s) on top of base */
fbi->lccr0 |= LCCR0_OUC;
pr_info("PXA Overlay driver loaded successfully!\n");
return 0;
}
static void __devexit pxafb_overlay_exit(struct pxafb_info *fbi)
{
int i;
if (!pxafb_overlay_supported())
return;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
for (i = 0; i < 2; i++)
unregister_framebuffer(&fbi->overlay[i].fb);
}
#else
static inline void pxafb_overlay_init(struct pxafb_info *fbi) {}
static inline void pxafb_overlay_exit(struct pxafb_info *fbi) {}
#endif /* CONFIG_FB_PXA_OVERLAY */
/*
* Calculate the PCD value from the clock rate (in picoseconds).
* We take account of the PPCR clock setting.
* From PXA Developer's Manual:
*
* PixelClock = LCLK
* -------------
* 2 ( PCD + 1 )
*
* PCD = LCLK
* ------------- - 1
* 2(PixelClock)
*
* Where:
* LCLK = LCD/Memory Clock
* PCD = LCCR3[7:0]
*
* PixelClock here is in Hz while the pixclock argument given is the
* period in picoseconds. Hence PixelClock = 1 / ( pixclock * 10^-12 )
*
* The function get_lclk_frequency_10khz returns LCLK in units of
* 10khz. Calling the result of this function lclk gives us the
* following
*
* PCD = (lclk * 10^4 ) * ( pixclock * 10^-12 )
* -------------------------------------- - 1
* 2
*
* Factoring the 10^4 and 10^-12 out gives 10^-8 == 1 / 100000000 as used below.
*/
static inline unsigned int get_pcd(struct pxafb_info *fbi,
unsigned int pixclock)
{
unsigned long long pcd;
/* FIXME: Need to take into account Double Pixel Clock mode
* (DPC) bit? or perhaps set it based on the various clock
* speeds */
pcd = (unsigned long long)(clk_get_rate(fbi->clk) / 10000);
pcd *= pixclock;
do_div(pcd, 100000000 * 2);
/* no need for this, since we should subtract 1 anyway. they cancel */
/* pcd += 1; */ /* make up for integer math truncations */
return (unsigned int)pcd;
}
/*
* Some touchscreens need hsync information from the video driver to
* function correctly. We export it here. Note that 'hsync_time' and
* the value returned from pxafb_get_hsync_time() is the *reciprocal*
* of the hsync period in seconds.
*/
static inline void set_hsync_time(struct pxafb_info *fbi, unsigned int pcd)
{
unsigned long htime;
if ((pcd == 0) || (fbi->fb.var.hsync_len == 0)) {
fbi->hsync_time = 0;
return;
}
htime = clk_get_rate(fbi->clk) / (pcd * fbi->fb.var.hsync_len);
fbi->hsync_time = htime;
}
unsigned long pxafb_get_hsync_time(struct device *dev)
{
struct pxafb_info *fbi = dev_get_drvdata(dev);
/* If display is blanked/suspended, hsync isn't active */
if (!fbi || (fbi->state != C_ENABLE))
return 0;
return fbi->hsync_time;
}
EXPORT_SYMBOL(pxafb_get_hsync_time);
static int setup_frame_dma(struct pxafb_info *fbi, int dma, int pal,
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
unsigned long start, size_t size)
{
struct pxafb_dma_descriptor *dma_desc, *pal_desc;
unsigned int dma_desc_off, pal_desc_off;
if (dma < 0 || dma >= DMA_MAX * 2)
return -EINVAL;
dma_desc = &fbi->dma_buff->dma_desc[dma];
dma_desc_off = offsetof(struct pxafb_dma_buff, dma_desc[dma]);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
dma_desc->fsadr = start;
dma_desc->fidr = 0;
dma_desc->ldcmd = size;
if (pal < 0 || pal >= PAL_MAX * 2) {
dma_desc->fdadr = fbi->dma_buff_phys + dma_desc_off;
fbi->fdadr[dma] = fbi->dma_buff_phys + dma_desc_off;
} else {
pal_desc = &fbi->dma_buff->pal_desc[pal];
pal_desc_off = offsetof(struct pxafb_dma_buff, pal_desc[pal]);
pal_desc->fsadr = fbi->dma_buff_phys + pal * PALETTE_SIZE;
pal_desc->fidr = 0;
if ((fbi->lccr4 & LCCR4_PAL_FOR_MASK) == LCCR4_PAL_FOR_0)
pal_desc->ldcmd = fbi->palette_size * sizeof(u16);
else
pal_desc->ldcmd = fbi->palette_size * sizeof(u32);
pal_desc->ldcmd |= LDCMD_PAL;
/* flip back and forth between palette and frame buffer */
pal_desc->fdadr = fbi->dma_buff_phys + dma_desc_off;
dma_desc->fdadr = fbi->dma_buff_phys + pal_desc_off;
fbi->fdadr[dma] = fbi->dma_buff_phys + dma_desc_off;
}
return 0;
}
static void setup_base_frame(struct pxafb_info *fbi,
struct fb_var_screeninfo *var,
int branch)
{
struct fb_fix_screeninfo *fix = &fbi->fb.fix;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
int nbytes, dma, pal, bpp = var->bits_per_pixel;
unsigned long offset;
dma = DMA_BASE + (branch ? DMA_MAX : 0);
pal = (bpp >= 16) ? PAL_NONE : PAL_BASE + (branch ? PAL_MAX : 0);
nbytes = fix->line_length * var->yres;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
offset = fix->line_length * var->yoffset + fbi->video_mem_phys;
if (fbi->lccr0 & LCCR0_SDS) {
nbytes = nbytes / 2;
setup_frame_dma(fbi, dma + 1, PAL_NONE, offset + nbytes, nbytes);
}
setup_frame_dma(fbi, dma, pal, offset, nbytes);
}
#ifdef CONFIG_FB_PXA_SMARTPANEL
static int setup_smart_dma(struct pxafb_info *fbi)
{
struct pxafb_dma_descriptor *dma_desc;
unsigned long dma_desc_off, cmd_buff_off;
dma_desc = &fbi->dma_buff->dma_desc[DMA_CMD];
dma_desc_off = offsetof(struct pxafb_dma_buff, dma_desc[DMA_CMD]);
cmd_buff_off = offsetof(struct pxafb_dma_buff, cmd_buff);
dma_desc->fdadr = fbi->dma_buff_phys + dma_desc_off;
dma_desc->fsadr = fbi->dma_buff_phys + cmd_buff_off;
dma_desc->fidr = 0;
dma_desc->ldcmd = fbi->n_smart_cmds * sizeof(uint16_t);
fbi->fdadr[DMA_CMD] = dma_desc->fdadr;
return 0;
}
int pxafb_smart_flush(struct fb_info *info)
{
struct pxafb_info *fbi = container_of(info, struct pxafb_info, fb);
uint32_t prsr;
int ret = 0;
/* disable controller until all registers are set up */
lcd_writel(fbi, LCCR0, fbi->reg_lccr0 & ~LCCR0_ENB);
/* 1. make it an even number of commands to align on 32-bit boundary
* 2. add the interrupt command to the end of the chain so we can
* keep track of the end of the transfer
*/
while (fbi->n_smart_cmds & 1)
fbi->smart_cmds[fbi->n_smart_cmds++] = SMART_CMD_NOOP;
fbi->smart_cmds[fbi->n_smart_cmds++] = SMART_CMD_INTERRUPT;
fbi->smart_cmds[fbi->n_smart_cmds++] = SMART_CMD_WAIT_FOR_VSYNC;
setup_smart_dma(fbi);
/* continue to execute next command */
prsr = lcd_readl(fbi, PRSR) | PRSR_ST_OK | PRSR_CON_NT;
lcd_writel(fbi, PRSR, prsr);
/* stop the processor in case it executed "wait for sync" cmd */
lcd_writel(fbi, CMDCR, 0x0001);
/* don't send interrupts for fifo underruns on channel 6 */
lcd_writel(fbi, LCCR5, LCCR5_IUM(6));
lcd_writel(fbi, LCCR1, fbi->reg_lccr1);
lcd_writel(fbi, LCCR2, fbi->reg_lccr2);
lcd_writel(fbi, LCCR3, fbi->reg_lccr3);
lcd_writel(fbi, LCCR4, fbi->reg_lccr4);
lcd_writel(fbi, FDADR0, fbi->fdadr[0]);
lcd_writel(fbi, FDADR6, fbi->fdadr[6]);
/* begin sending */
lcd_writel(fbi, LCCR0, fbi->reg_lccr0 | LCCR0_ENB);
if (wait_for_completion_timeout(&fbi->command_done, HZ/2) == 0) {
pr_warning("%s: timeout waiting for command done\n",
__func__);
ret = -ETIMEDOUT;
}
/* quick disable */
prsr = lcd_readl(fbi, PRSR) & ~(PRSR_ST_OK | PRSR_CON_NT);
lcd_writel(fbi, PRSR, prsr);
lcd_writel(fbi, LCCR0, fbi->reg_lccr0 & ~LCCR0_ENB);
lcd_writel(fbi, FDADR6, 0);
fbi->n_smart_cmds = 0;
return ret;
}
int pxafb_smart_queue(struct fb_info *info, uint16_t *cmds, int n_cmds)
{
int i;
struct pxafb_info *fbi = container_of(info, struct pxafb_info, fb);
for (i = 0; i < n_cmds; i++, cmds++) {
/* if it is a software delay, flush and delay */
if ((*cmds & 0xff00) == SMART_CMD_DELAY) {
pxafb_smart_flush(info);
mdelay(*cmds & 0xff);
continue;
}
/* leave 2 commands for INTERRUPT and WAIT_FOR_SYNC */
if (fbi->n_smart_cmds == CMD_BUFF_SIZE - 8)
pxafb_smart_flush(info);
fbi->smart_cmds[fbi->n_smart_cmds++] = *cmds;
}
return 0;
}
static unsigned int __smart_timing(unsigned time_ns, unsigned long lcd_clk)
{
unsigned int t = (time_ns * (lcd_clk / 1000000) / 1000);
return (t == 0) ? 1 : t;
}
static void setup_smart_timing(struct pxafb_info *fbi,
struct fb_var_screeninfo *var)
{
struct pxafb_mach_info *inf = fbi->dev->platform_data;
struct pxafb_mode_info *mode = &inf->modes[0];
unsigned long lclk = clk_get_rate(fbi->clk);
unsigned t1, t2, t3, t4;
t1 = max(mode->a0csrd_set_hld, mode->a0cswr_set_hld);
t2 = max(mode->rd_pulse_width, mode->wr_pulse_width);
t3 = mode->op_hold_time;
t4 = mode->cmd_inh_time;
fbi->reg_lccr1 =
LCCR1_DisWdth(var->xres) |
LCCR1_BegLnDel(__smart_timing(t1, lclk)) |
LCCR1_EndLnDel(__smart_timing(t2, lclk)) |
LCCR1_HorSnchWdth(__smart_timing(t3, lclk));
fbi->reg_lccr2 = LCCR2_DisHght(var->yres);
fbi->reg_lccr3 = fbi->lccr3 | LCCR3_PixClkDiv(__smart_timing(t4, lclk));
fbi->reg_lccr3 |= (var->sync & FB_SYNC_HOR_HIGH_ACT) ? LCCR3_HSP : 0;
fbi->reg_lccr3 |= (var->sync & FB_SYNC_VERT_HIGH_ACT) ? LCCR3_VSP : 0;
/* FIXME: make this configurable */
fbi->reg_cmdcr = 1;
}
static int pxafb_smart_thread(void *arg)
{
struct pxafb_info *fbi = arg;
struct pxafb_mach_info *inf = fbi->dev->platform_data;
if (!inf->smart_update) {
pr_err("%s: not properly initialized, thread terminated\n",
__func__);
return -EINVAL;
}
inf = fbi->dev->platform_data;
pr_debug("%s(): task starting\n", __func__);
set_freezable();
while (!kthread_should_stop()) {
if (try_to_freeze())
continue;
mutex_lock(&fbi->ctrlr_lock);
if (fbi->state == C_ENABLE) {
inf->smart_update(&fbi->fb);
complete(&fbi->refresh_done);
}
mutex_unlock(&fbi->ctrlr_lock);
set_current_state(TASK_INTERRUPTIBLE);
schedule_timeout(30 * HZ / 1000);
}
pr_debug("%s(): task ending\n", __func__);
return 0;
}
static int pxafb_smart_init(struct pxafb_info *fbi)
{
if (!(fbi->lccr0 & LCCR0_LCDT))
return 0;
fbi->smart_cmds = (uint16_t *) fbi->dma_buff->cmd_buff;
fbi->n_smart_cmds = 0;
init_completion(&fbi->command_done);
init_completion(&fbi->refresh_done);
fbi->smart_thread = kthread_run(pxafb_smart_thread, fbi,
"lcd_refresh");
if (IS_ERR(fbi->smart_thread)) {
pr_err("%s: unable to create kernel thread\n", __func__);
return PTR_ERR(fbi->smart_thread);
}
return 0;
}
#else
int pxafb_smart_queue(struct fb_info *info, uint16_t *cmds, int n_cmds)
{
return 0;
}
int pxafb_smart_flush(struct fb_info *info)
{
return 0;
}
static inline int pxafb_smart_init(struct pxafb_info *fbi) { return 0; }
#endif /* CONFIG_FB_PXA_SMARTPANEL */
static void setup_parallel_timing(struct pxafb_info *fbi,
struct fb_var_screeninfo *var)
{
unsigned int lines_per_panel, pcd = get_pcd(fbi, var->pixclock);
fbi->reg_lccr1 =
LCCR1_DisWdth(var->xres) +
LCCR1_HorSnchWdth(var->hsync_len) +
LCCR1_BegLnDel(var->left_margin) +
LCCR1_EndLnDel(var->right_margin);
/*
* If we have a dual scan LCD, we need to halve
* the YRES parameter.
*/
lines_per_panel = var->yres;
if ((fbi->lccr0 & LCCR0_SDS) == LCCR0_Dual)
lines_per_panel /= 2;
fbi->reg_lccr2 =
LCCR2_DisHght(lines_per_panel) +
LCCR2_VrtSnchWdth(var->vsync_len) +
LCCR2_BegFrmDel(var->upper_margin) +
LCCR2_EndFrmDel(var->lower_margin);
fbi->reg_lccr3 = fbi->lccr3 |
(var->sync & FB_SYNC_HOR_HIGH_ACT ?
LCCR3_HorSnchH : LCCR3_HorSnchL) |
(var->sync & FB_SYNC_VERT_HIGH_ACT ?
LCCR3_VrtSnchH : LCCR3_VrtSnchL);
if (pcd) {
fbi->reg_lccr3 |= LCCR3_PixClkDiv(pcd);
set_hsync_time(fbi, pcd);
}
}
/*
* pxafb_activate_var():
* Configures LCD Controller based on entries in var parameter.
* Settings are only written to the controller if changes were made.
*/
static int pxafb_activate_var(struct fb_var_screeninfo *var,
struct pxafb_info *fbi)
{
u_long flags;
/* Update shadow copy atomically */
local_irq_save(flags);
#ifdef CONFIG_FB_PXA_SMARTPANEL
if (fbi->lccr0 & LCCR0_LCDT)
setup_smart_timing(fbi, var);
else
#endif
setup_parallel_timing(fbi, var);
setup_base_frame(fbi, var, 0);
fbi->reg_lccr0 = fbi->lccr0 |
(LCCR0_LDM | LCCR0_SFM | LCCR0_IUM | LCCR0_EFM |
LCCR0_QDM | LCCR0_BM | LCCR0_OUM);
fbi->reg_lccr3 |= pxafb_var_to_lccr3(var);
fbi->reg_lccr4 = lcd_readl(fbi, LCCR4) & ~LCCR4_PAL_FOR_MASK;
fbi->reg_lccr4 |= (fbi->lccr4 & LCCR4_PAL_FOR_MASK);
local_irq_restore(flags);
/*
* Only update the registers if the controller is enabled
* and something has changed.
*/
if ((lcd_readl(fbi, LCCR0) != fbi->reg_lccr0) ||
(lcd_readl(fbi, LCCR1) != fbi->reg_lccr1) ||
(lcd_readl(fbi, LCCR2) != fbi->reg_lccr2) ||
(lcd_readl(fbi, LCCR3) != fbi->reg_lccr3) ||
(lcd_readl(fbi, LCCR4) != fbi->reg_lccr4) ||
(lcd_readl(fbi, FDADR0) != fbi->fdadr[0]) ||
(lcd_readl(fbi, FDADR1) != fbi->fdadr[1]))
pxafb_schedule_work(fbi, C_REENABLE);
return 0;
}
/*
* NOTE! The following functions are purely helpers for set_ctrlr_state.
* Do not call them directly; set_ctrlr_state does the correct serialisation
* to ensure that things happen in the right way 100% of time time.
* -- rmk
*/
static inline void __pxafb_backlight_power(struct pxafb_info *fbi, int on)
{
pr_debug("pxafb: backlight o%s\n", on ? "n" : "ff");
if (fbi->backlight_power)
fbi->backlight_power(on);
}
static inline void __pxafb_lcd_power(struct pxafb_info *fbi, int on)
{
pr_debug("pxafb: LCD power o%s\n", on ? "n" : "ff");
if (fbi->lcd_power)
fbi->lcd_power(on, &fbi->fb.var);
}
static void pxafb_enable_controller(struct pxafb_info *fbi)
{
pr_debug("pxafb: Enabling LCD controller\n");
pr_debug("fdadr0 0x%08x\n", (unsigned int) fbi->fdadr[0]);
pr_debug("fdadr1 0x%08x\n", (unsigned int) fbi->fdadr[1]);
pr_debug("reg_lccr0 0x%08x\n", (unsigned int) fbi->reg_lccr0);
pr_debug("reg_lccr1 0x%08x\n", (unsigned int) fbi->reg_lccr1);
pr_debug("reg_lccr2 0x%08x\n", (unsigned int) fbi->reg_lccr2);
pr_debug("reg_lccr3 0x%08x\n", (unsigned int) fbi->reg_lccr3);
/* enable LCD controller clock */
clk_enable(fbi->clk);
if (fbi->lccr0 & LCCR0_LCDT)
return;
/* Sequence from 11.7.10 */
lcd_writel(fbi, LCCR4, fbi->reg_lccr4);
lcd_writel(fbi, LCCR3, fbi->reg_lccr3);
lcd_writel(fbi, LCCR2, fbi->reg_lccr2);
lcd_writel(fbi, LCCR1, fbi->reg_lccr1);
lcd_writel(fbi, LCCR0, fbi->reg_lccr0 & ~LCCR0_ENB);
lcd_writel(fbi, FDADR0, fbi->fdadr[0]);
lcd_writel(fbi, FDADR1, fbi->fdadr[1]);
lcd_writel(fbi, LCCR0, fbi->reg_lccr0 | LCCR0_ENB);
}
static void pxafb_disable_controller(struct pxafb_info *fbi)
{
uint32_t lccr0;
#ifdef CONFIG_FB_PXA_SMARTPANEL
if (fbi->lccr0 & LCCR0_LCDT) {
wait_for_completion_timeout(&fbi->refresh_done,
200 * HZ / 1000);
return;
}
#endif
/* Clear LCD Status Register */
lcd_writel(fbi, LCSR, 0xffffffff);
lccr0 = lcd_readl(fbi, LCCR0) & ~LCCR0_LDM;
lcd_writel(fbi, LCCR0, lccr0);
lcd_writel(fbi, LCCR0, lccr0 | LCCR0_DIS);
wait_for_completion_timeout(&fbi->disable_done, 200 * HZ / 1000);
/* disable LCD controller clock */
clk_disable(fbi->clk);
}
/*
* pxafb_handle_irq: Handle 'LCD DONE' interrupts.
*/
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 21:55:46 +08:00
static irqreturn_t pxafb_handle_irq(int irq, void *dev_id)
{
struct pxafb_info *fbi = dev_id;
unsigned int lccr0, lcsr;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
lcsr = lcd_readl(fbi, LCSR);
if (lcsr & LCSR_LDD) {
lccr0 = lcd_readl(fbi, LCCR0);
lcd_writel(fbi, LCCR0, lccr0 | LCCR0_LDM);
complete(&fbi->disable_done);
}
#ifdef CONFIG_FB_PXA_SMARTPANEL
if (lcsr & LCSR_CMD_INT)
complete(&fbi->command_done);
#endif
lcd_writel(fbi, LCSR, lcsr);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
#ifdef CONFIG_FB_PXA_OVERLAY
{
unsigned int lcsr1 = lcd_readl(fbi, LCSR1);
if (lcsr1 & LCSR1_BS(1))
complete(&fbi->overlay[0].branch_done);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
if (lcsr1 & LCSR1_BS(2))
complete(&fbi->overlay[1].branch_done);
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
lcd_writel(fbi, LCSR1, lcsr1);
}
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
#endif
return IRQ_HANDLED;
}
/*
* This function must be called from task context only, since it will
* sleep when disabling the LCD controller, or if we get two contending
* processes trying to alter state.
*/
static void set_ctrlr_state(struct pxafb_info *fbi, u_int state)
{
u_int old_state;
mutex_lock(&fbi->ctrlr_lock);
old_state = fbi->state;
/*
* Hack around fbcon initialisation.
*/
if (old_state == C_STARTUP && state == C_REENABLE)
state = C_ENABLE;
switch (state) {
case C_DISABLE_CLKCHANGE:
/*
* Disable controller for clock change. If the
* controller is already disabled, then do nothing.
*/
if (old_state != C_DISABLE && old_state != C_DISABLE_PM) {
fbi->state = state;
/* TODO __pxafb_lcd_power(fbi, 0); */
pxafb_disable_controller(fbi);
}
break;
case C_DISABLE_PM:
case C_DISABLE:
/*
* Disable controller
*/
if (old_state != C_DISABLE) {
fbi->state = state;
__pxafb_backlight_power(fbi, 0);
__pxafb_lcd_power(fbi, 0);
if (old_state != C_DISABLE_CLKCHANGE)
pxafb_disable_controller(fbi);
}
break;
case C_ENABLE_CLKCHANGE:
/*
* Enable the controller after clock change. Only
* do this if we were disabled for the clock change.
*/
if (old_state == C_DISABLE_CLKCHANGE) {
fbi->state = C_ENABLE;
pxafb_enable_controller(fbi);
/* TODO __pxafb_lcd_power(fbi, 1); */
}
break;
case C_REENABLE:
/*
* Re-enable the controller only if it was already
* enabled. This is so we reprogram the control
* registers.
*/
if (old_state == C_ENABLE) {
__pxafb_lcd_power(fbi, 0);
pxafb_disable_controller(fbi);
pxafb_enable_controller(fbi);
__pxafb_lcd_power(fbi, 1);
}
break;
case C_ENABLE_PM:
/*
* Re-enable the controller after PM. This is not
* perfect - think about the case where we were doing
* a clock change, and we suspended half-way through.
*/
if (old_state != C_DISABLE_PM)
break;
/* fall through */
case C_ENABLE:
/*
* Power up the LCD screen, enable controller, and
* turn on the backlight.
*/
if (old_state != C_ENABLE) {
fbi->state = C_ENABLE;
pxafb_enable_controller(fbi);
__pxafb_lcd_power(fbi, 1);
__pxafb_backlight_power(fbi, 1);
}
break;
}
mutex_unlock(&fbi->ctrlr_lock);
}
/*
* Our LCD controller task (which is called when we blank or unblank)
* via keventd.
*/
static void pxafb_task(struct work_struct *work)
{
struct pxafb_info *fbi =
container_of(work, struct pxafb_info, task);
u_int state = xchg(&fbi->task_state, -1);
set_ctrlr_state(fbi, state);
}
#ifdef CONFIG_CPU_FREQ
/*
* CPU clock speed change handler. We need to adjust the LCD timing
* parameters when the CPU clock is adjusted by the power management
* subsystem.
*
* TODO: Determine why f->new != 10*get_lclk_frequency_10khz()
*/
static int
pxafb_freq_transition(struct notifier_block *nb, unsigned long val, void *data)
{
struct pxafb_info *fbi = TO_INF(nb, freq_transition);
/* TODO struct cpufreq_freqs *f = data; */
u_int pcd;
switch (val) {
case CPUFREQ_PRECHANGE:
set_ctrlr_state(fbi, C_DISABLE_CLKCHANGE);
break;
case CPUFREQ_POSTCHANGE:
pcd = get_pcd(fbi, fbi->fb.var.pixclock);
set_hsync_time(fbi, pcd);
fbi->reg_lccr3 = (fbi->reg_lccr3 & ~0xff) |
LCCR3_PixClkDiv(pcd);
set_ctrlr_state(fbi, C_ENABLE_CLKCHANGE);
break;
}
return 0;
}
static int
pxafb_freq_policy(struct notifier_block *nb, unsigned long val, void *data)
{
struct pxafb_info *fbi = TO_INF(nb, freq_policy);
struct fb_var_screeninfo *var = &fbi->fb.var;
struct cpufreq_policy *policy = data;
switch (val) {
case CPUFREQ_ADJUST:
case CPUFREQ_INCOMPATIBLE:
pr_debug("min dma period: %d ps, "
"new clock %d kHz\n", pxafb_display_dma_period(var),
policy->max);
/* TODO: fill in min/max values */
break;
}
return 0;
}
#endif
#ifdef CONFIG_PM
/*
* Power management hooks. Note that we won't be called from IRQ context,
* unlike the blank functions above, so we may sleep.
*/
static int pxafb_suspend(struct device *dev)
{
struct pxafb_info *fbi = dev_get_drvdata(dev);
set_ctrlr_state(fbi, C_DISABLE_PM);
return 0;
}
static int pxafb_resume(struct device *dev)
{
struct pxafb_info *fbi = dev_get_drvdata(dev);
set_ctrlr_state(fbi, C_ENABLE_PM);
return 0;
}
static const struct dev_pm_ops pxafb_pm_ops = {
.suspend = pxafb_suspend,
.resume = pxafb_resume,
};
#endif
static int __devinit pxafb_init_video_memory(struct pxafb_info *fbi)
{
int size = PAGE_ALIGN(fbi->video_mem_size);
fbi->video_mem = alloc_pages_exact(size, GFP_KERNEL | __GFP_ZERO);
if (fbi->video_mem == NULL)
return -ENOMEM;
fbi->video_mem_phys = virt_to_phys(fbi->video_mem);
fbi->video_mem_size = size;
fbi->fb.fix.smem_start = fbi->video_mem_phys;
fbi->fb.fix.smem_len = fbi->video_mem_size;
fbi->fb.screen_base = fbi->video_mem;
return fbi->video_mem ? 0 : -ENOMEM;
}
static void pxafb_decode_mach_info(struct pxafb_info *fbi,
struct pxafb_mach_info *inf)
{
unsigned int lcd_conn = inf->lcd_conn;
struct pxafb_mode_info *m;
int i;
fbi->cmap_inverse = inf->cmap_inverse;
fbi->cmap_static = inf->cmap_static;
fbi->lccr4 = inf->lccr4;
switch (lcd_conn & LCD_TYPE_MASK) {
case LCD_TYPE_MONO_STN:
fbi->lccr0 = LCCR0_CMS;
break;
case LCD_TYPE_MONO_DSTN:
fbi->lccr0 = LCCR0_CMS | LCCR0_SDS;
break;
case LCD_TYPE_COLOR_STN:
fbi->lccr0 = 0;
break;
case LCD_TYPE_COLOR_DSTN:
fbi->lccr0 = LCCR0_SDS;
break;
case LCD_TYPE_COLOR_TFT:
fbi->lccr0 = LCCR0_PAS;
break;
case LCD_TYPE_SMART_PANEL:
fbi->lccr0 = LCCR0_LCDT | LCCR0_PAS;
break;
default:
/* fall back to backward compatibility way */
fbi->lccr0 = inf->lccr0;
fbi->lccr3 = inf->lccr3;
goto decode_mode;
}
if (lcd_conn == LCD_MONO_STN_8BPP)
fbi->lccr0 |= LCCR0_DPD;
fbi->lccr0 |= (lcd_conn & LCD_ALTERNATE_MAPPING) ? LCCR0_LDDALT : 0;
fbi->lccr3 = LCCR3_Acb((inf->lcd_conn >> 10) & 0xff);
fbi->lccr3 |= (lcd_conn & LCD_BIAS_ACTIVE_LOW) ? LCCR3_OEP : 0;
fbi->lccr3 |= (lcd_conn & LCD_PCLK_EDGE_FALL) ? LCCR3_PCP : 0;
decode_mode:
pxafb_setmode(&fbi->fb.var, &inf->modes[0]);
/* decide video memory size as follows:
* 1. default to mode of maximum resolution
* 2. allow platform to override
* 3. allow module parameter to override
*/
for (i = 0, m = &inf->modes[0]; i < inf->num_modes; i++, m++)
fbi->video_mem_size = max_t(size_t, fbi->video_mem_size,
m->xres * m->yres * m->bpp / 8);
if (inf->video_mem_size > fbi->video_mem_size)
fbi->video_mem_size = inf->video_mem_size;
if (video_mem_size > fbi->video_mem_size)
fbi->video_mem_size = video_mem_size;
}
static struct pxafb_info * __devinit pxafb_init_fbinfo(struct device *dev)
{
struct pxafb_info *fbi;
void *addr;
struct pxafb_mach_info *inf = dev->platform_data;
/* Alloc the pxafb_info and pseudo_palette in one step */
fbi = kmalloc(sizeof(struct pxafb_info) + sizeof(u32) * 16, GFP_KERNEL);
if (!fbi)
return NULL;
memset(fbi, 0, sizeof(struct pxafb_info));
fbi->dev = dev;
fbi->clk = clk_get(dev, NULL);
if (IS_ERR(fbi->clk)) {
kfree(fbi);
return NULL;
}
strcpy(fbi->fb.fix.id, PXA_NAME);
fbi->fb.fix.type = FB_TYPE_PACKED_PIXELS;
fbi->fb.fix.type_aux = 0;
fbi->fb.fix.xpanstep = 0;
fbi->fb.fix.ypanstep = 1;
fbi->fb.fix.ywrapstep = 0;
fbi->fb.fix.accel = FB_ACCEL_NONE;
fbi->fb.var.nonstd = 0;
fbi->fb.var.activate = FB_ACTIVATE_NOW;
fbi->fb.var.height = -1;
fbi->fb.var.width = -1;
fbi->fb.var.accel_flags = FB_ACCELF_TEXT;
fbi->fb.var.vmode = FB_VMODE_NONINTERLACED;
fbi->fb.fbops = &pxafb_ops;
fbi->fb.flags = FBINFO_DEFAULT;
fbi->fb.node = -1;
addr = fbi;
addr = addr + sizeof(struct pxafb_info);
fbi->fb.pseudo_palette = addr;
fbi->state = C_STARTUP;
fbi->task_state = (u_char)-1;
pxafb_decode_mach_info(fbi, inf);
init_waitqueue_head(&fbi->ctrlr_wait);
INIT_WORK(&fbi->task, pxafb_task);
mutex_init(&fbi->ctrlr_lock);
init_completion(&fbi->disable_done);
return fbi;
}
#ifdef CONFIG_FB_PXA_PARAMETERS
static int __devinit parse_opt_mode(struct device *dev, const char *this_opt)
{
struct pxafb_mach_info *inf = dev->platform_data;
const char *name = this_opt+5;
unsigned int namelen = strlen(name);
int res_specified = 0, bpp_specified = 0;
unsigned int xres = 0, yres = 0, bpp = 0;
int yres_specified = 0;
int i;
for (i = namelen-1; i >= 0; i--) {
switch (name[i]) {
case '-':
namelen = i;
if (!bpp_specified && !yres_specified) {
bpp = simple_strtoul(&name[i+1], NULL, 0);
bpp_specified = 1;
} else
goto done;
break;
case 'x':
if (!yres_specified) {
yres = simple_strtoul(&name[i+1], NULL, 0);
yres_specified = 1;
} else
goto done;
break;
case '0' ... '9':
break;
default:
goto done;
}
}
if (i < 0 && yres_specified) {
xres = simple_strtoul(name, NULL, 0);
res_specified = 1;
}
done:
if (res_specified) {
dev_info(dev, "overriding resolution: %dx%d\n", xres, yres);
inf->modes[0].xres = xres; inf->modes[0].yres = yres;
}
if (bpp_specified)
switch (bpp) {
case 1:
case 2:
case 4:
case 8:
case 16:
inf->modes[0].bpp = bpp;
dev_info(dev, "overriding bit depth: %d\n", bpp);
break;
default:
dev_err(dev, "Depth %d is not valid\n", bpp);
return -EINVAL;
}
return 0;
}
static int __devinit parse_opt(struct device *dev, char *this_opt)
{
struct pxafb_mach_info *inf = dev->platform_data;
struct pxafb_mode_info *mode = &inf->modes[0];
char s[64];
s[0] = '\0';
if (!strncmp(this_opt, "vmem:", 5)) {
video_mem_size = memparse(this_opt + 5, NULL);
} else if (!strncmp(this_opt, "mode:", 5)) {
return parse_opt_mode(dev, this_opt);
} else if (!strncmp(this_opt, "pixclock:", 9)) {
mode->pixclock = simple_strtoul(this_opt+9, NULL, 0);
sprintf(s, "pixclock: %ld\n", mode->pixclock);
} else if (!strncmp(this_opt, "left:", 5)) {
mode->left_margin = simple_strtoul(this_opt+5, NULL, 0);
sprintf(s, "left: %u\n", mode->left_margin);
} else if (!strncmp(this_opt, "right:", 6)) {
mode->right_margin = simple_strtoul(this_opt+6, NULL, 0);
sprintf(s, "right: %u\n", mode->right_margin);
} else if (!strncmp(this_opt, "upper:", 6)) {
mode->upper_margin = simple_strtoul(this_opt+6, NULL, 0);
sprintf(s, "upper: %u\n", mode->upper_margin);
} else if (!strncmp(this_opt, "lower:", 6)) {
mode->lower_margin = simple_strtoul(this_opt+6, NULL, 0);
sprintf(s, "lower: %u\n", mode->lower_margin);
} else if (!strncmp(this_opt, "hsynclen:", 9)) {
mode->hsync_len = simple_strtoul(this_opt+9, NULL, 0);
sprintf(s, "hsynclen: %u\n", mode->hsync_len);
} else if (!strncmp(this_opt, "vsynclen:", 9)) {
mode->vsync_len = simple_strtoul(this_opt+9, NULL, 0);
sprintf(s, "vsynclen: %u\n", mode->vsync_len);
} else if (!strncmp(this_opt, "hsync:", 6)) {
if (simple_strtoul(this_opt+6, NULL, 0) == 0) {
sprintf(s, "hsync: Active Low\n");
mode->sync &= ~FB_SYNC_HOR_HIGH_ACT;
} else {
sprintf(s, "hsync: Active High\n");
mode->sync |= FB_SYNC_HOR_HIGH_ACT;
}
} else if (!strncmp(this_opt, "vsync:", 6)) {
if (simple_strtoul(this_opt+6, NULL, 0) == 0) {
sprintf(s, "vsync: Active Low\n");
mode->sync &= ~FB_SYNC_VERT_HIGH_ACT;
} else {
sprintf(s, "vsync: Active High\n");
mode->sync |= FB_SYNC_VERT_HIGH_ACT;
}
} else if (!strncmp(this_opt, "dpc:", 4)) {
if (simple_strtoul(this_opt+4, NULL, 0) == 0) {
sprintf(s, "double pixel clock: false\n");
inf->lccr3 &= ~LCCR3_DPC;
} else {
sprintf(s, "double pixel clock: true\n");
inf->lccr3 |= LCCR3_DPC;
}
} else if (!strncmp(this_opt, "outputen:", 9)) {
if (simple_strtoul(this_opt+9, NULL, 0) == 0) {
sprintf(s, "output enable: active low\n");
inf->lccr3 = (inf->lccr3 & ~LCCR3_OEP) | LCCR3_OutEnL;
} else {
sprintf(s, "output enable: active high\n");
inf->lccr3 = (inf->lccr3 & ~LCCR3_OEP) | LCCR3_OutEnH;
}
} else if (!strncmp(this_opt, "pixclockpol:", 12)) {
if (simple_strtoul(this_opt+12, NULL, 0) == 0) {
sprintf(s, "pixel clock polarity: falling edge\n");
inf->lccr3 = (inf->lccr3 & ~LCCR3_PCP) | LCCR3_PixFlEdg;
} else {
sprintf(s, "pixel clock polarity: rising edge\n");
inf->lccr3 = (inf->lccr3 & ~LCCR3_PCP) | LCCR3_PixRsEdg;
}
} else if (!strncmp(this_opt, "color", 5)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_CMS) | LCCR0_Color;
} else if (!strncmp(this_opt, "mono", 4)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_CMS) | LCCR0_Mono;
} else if (!strncmp(this_opt, "active", 6)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_PAS) | LCCR0_Act;
} else if (!strncmp(this_opt, "passive", 7)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_PAS) | LCCR0_Pas;
} else if (!strncmp(this_opt, "single", 6)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_SDS) | LCCR0_Sngl;
} else if (!strncmp(this_opt, "dual", 4)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_SDS) | LCCR0_Dual;
} else if (!strncmp(this_opt, "4pix", 4)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_DPD) | LCCR0_4PixMono;
} else if (!strncmp(this_opt, "8pix", 4)) {
inf->lccr0 = (inf->lccr0 & ~LCCR0_DPD) | LCCR0_8PixMono;
} else {
dev_err(dev, "unknown option: %s\n", this_opt);
return -EINVAL;
}
if (s[0] != '\0')
dev_info(dev, "override %s", s);
return 0;
}
static int __devinit pxafb_parse_options(struct device *dev, char *options)
{
char *this_opt;
int ret;
if (!options || !*options)
return 0;
dev_dbg(dev, "options are \"%s\"\n", options ? options : "null");
/* could be made table driven or similar?... */
while ((this_opt = strsep(&options, ",")) != NULL) {
ret = parse_opt(dev, this_opt);
if (ret)
return ret;
}
return 0;
}
static char g_options[256] __devinitdata = "";
#ifndef MODULE
static int __init pxafb_setup_options(void)
{
char *options = NULL;
if (fb_get_options("pxafb", &options))
return -ENODEV;
if (options)
strlcpy(g_options, options, sizeof(g_options));
return 0;
}
#else
#define pxafb_setup_options() (0)
module_param_string(options, g_options, sizeof(g_options), 0);
MODULE_PARM_DESC(options, "LCD parameters (see Documentation/fb/pxafb.txt)");
#endif
#else
#define pxafb_parse_options(...) (0)
#define pxafb_setup_options() (0)
#endif
#ifdef DEBUG_VAR
/* Check for various illegal bit-combinations. Currently only
* a warning is given. */
static void __devinit pxafb_check_options(struct device *dev,
struct pxafb_mach_info *inf)
{
if (inf->lcd_conn)
return;
if (inf->lccr0 & LCCR0_INVALID_CONFIG_MASK)
dev_warn(dev, "machine LCCR0 setting contains "
"illegal bits: %08x\n",
inf->lccr0 & LCCR0_INVALID_CONFIG_MASK);
if (inf->lccr3 & LCCR3_INVALID_CONFIG_MASK)
dev_warn(dev, "machine LCCR3 setting contains "
"illegal bits: %08x\n",
inf->lccr3 & LCCR3_INVALID_CONFIG_MASK);
if (inf->lccr0 & LCCR0_DPD &&
((inf->lccr0 & LCCR0_PAS) != LCCR0_Pas ||
(inf->lccr0 & LCCR0_SDS) != LCCR0_Sngl ||
(inf->lccr0 & LCCR0_CMS) != LCCR0_Mono))
dev_warn(dev, "Double Pixel Data (DPD) mode is "
"only valid in passive mono"
" single panel mode\n");
if ((inf->lccr0 & LCCR0_PAS) == LCCR0_Act &&
(inf->lccr0 & LCCR0_SDS) == LCCR0_Dual)
dev_warn(dev, "Dual panel only valid in passive mode\n");
if ((inf->lccr0 & LCCR0_PAS) == LCCR0_Pas &&
(inf->modes->upper_margin || inf->modes->lower_margin))
dev_warn(dev, "Upper and lower margins must be 0 in "
"passive mode\n");
}
#else
#define pxafb_check_options(...) do {} while (0)
#endif
static int __devinit pxafb_probe(struct platform_device *dev)
{
struct pxafb_info *fbi;
struct pxafb_mach_info *inf;
struct resource *r;
int irq, ret;
dev_dbg(&dev->dev, "pxafb_probe\n");
inf = dev->dev.platform_data;
ret = -ENOMEM;
fbi = NULL;
if (!inf)
goto failed;
ret = pxafb_parse_options(&dev->dev, g_options);
if (ret < 0)
goto failed;
pxafb_check_options(&dev->dev, inf);
dev_dbg(&dev->dev, "got a %dx%dx%d LCD\n",
inf->modes->xres,
inf->modes->yres,
inf->modes->bpp);
if (inf->modes->xres == 0 ||
inf->modes->yres == 0 ||
inf->modes->bpp == 0) {
dev_err(&dev->dev, "Invalid resolution or bit depth\n");
ret = -EINVAL;
goto failed;
}
fbi = pxafb_init_fbinfo(&dev->dev);
if (!fbi) {
/* only reason for pxafb_init_fbinfo to fail is kmalloc */
dev_err(&dev->dev, "Failed to initialize framebuffer device\n");
ret = -ENOMEM;
goto failed;
}
if (cpu_is_pxa3xx() && inf->acceleration_enabled)
fbi->fb.fix.accel = FB_ACCEL_PXA3XX;
fbi->backlight_power = inf->pxafb_backlight_power;
fbi->lcd_power = inf->pxafb_lcd_power;
r = platform_get_resource(dev, IORESOURCE_MEM, 0);
if (r == NULL) {
dev_err(&dev->dev, "no I/O memory resource defined\n");
ret = -ENODEV;
goto failed_fbi;
}
r = request_mem_region(r->start, resource_size(r), dev->name);
if (r == NULL) {
dev_err(&dev->dev, "failed to request I/O memory\n");
ret = -EBUSY;
goto failed_fbi;
}
fbi->mmio_base = ioremap(r->start, resource_size(r));
if (fbi->mmio_base == NULL) {
dev_err(&dev->dev, "failed to map I/O memory\n");
ret = -EBUSY;
goto failed_free_res;
}
fbi->dma_buff_size = PAGE_ALIGN(sizeof(struct pxafb_dma_buff));
fbi->dma_buff = dma_alloc_coherent(fbi->dev, fbi->dma_buff_size,
&fbi->dma_buff_phys, GFP_KERNEL);
if (fbi->dma_buff == NULL) {
dev_err(&dev->dev, "failed to allocate memory for DMA\n");
ret = -ENOMEM;
goto failed_free_io;
}
ret = pxafb_init_video_memory(fbi);
if (ret) {
dev_err(&dev->dev, "Failed to allocate video RAM: %d\n", ret);
ret = -ENOMEM;
goto failed_free_dma;
}
irq = platform_get_irq(dev, 0);
if (irq < 0) {
dev_err(&dev->dev, "no IRQ defined\n");
ret = -ENODEV;
goto failed_free_mem;
}
ret = request_irq(irq, pxafb_handle_irq, IRQF_DISABLED, "LCD", fbi);
if (ret) {
dev_err(&dev->dev, "request_irq failed: %d\n", ret);
ret = -EBUSY;
goto failed_free_mem;
}
ret = pxafb_smart_init(fbi);
if (ret) {
dev_err(&dev->dev, "failed to initialize smartpanel\n");
goto failed_free_irq;
}
/*
* This makes sure that our colour bitfield
* descriptors are correctly initialised.
*/
ret = pxafb_check_var(&fbi->fb.var, &fbi->fb);
if (ret) {
dev_err(&dev->dev, "failed to get suitable mode\n");
goto failed_free_irq;
}
ret = pxafb_set_par(&fbi->fb);
if (ret) {
dev_err(&dev->dev, "Failed to set parameters\n");
goto failed_free_irq;
}
platform_set_drvdata(dev, fbi);
ret = register_framebuffer(&fbi->fb);
if (ret < 0) {
dev_err(&dev->dev,
"Failed to register framebuffer device: %d\n", ret);
goto failed_free_cmap;
}
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
pxafb_overlay_init(fbi);
#ifdef CONFIG_CPU_FREQ
fbi->freq_transition.notifier_call = pxafb_freq_transition;
fbi->freq_policy.notifier_call = pxafb_freq_policy;
cpufreq_register_notifier(&fbi->freq_transition,
CPUFREQ_TRANSITION_NOTIFIER);
cpufreq_register_notifier(&fbi->freq_policy,
CPUFREQ_POLICY_NOTIFIER);
#endif
/*
* Ok, now enable the LCD controller
*/
set_ctrlr_state(fbi, C_ENABLE);
return 0;
failed_free_cmap:
if (fbi->fb.cmap.len)
fb_dealloc_cmap(&fbi->fb.cmap);
failed_free_irq:
free_irq(irq, fbi);
failed_free_mem:
free_pages_exact(fbi->video_mem, fbi->video_mem_size);
failed_free_dma:
dma_free_coherent(&dev->dev, fbi->dma_buff_size,
fbi->dma_buff, fbi->dma_buff_phys);
failed_free_io:
iounmap(fbi->mmio_base);
failed_free_res:
release_mem_region(r->start, resource_size(r));
failed_fbi:
clk_put(fbi->clk);
platform_set_drvdata(dev, NULL);
kfree(fbi);
failed:
return ret;
}
static int __devexit pxafb_remove(struct platform_device *dev)
{
struct pxafb_info *fbi = platform_get_drvdata(dev);
struct resource *r;
int irq;
struct fb_info *info;
if (!fbi)
return 0;
info = &fbi->fb;
[ARM] pxafb: add support for overlay1 and overlay2 as framebuffer devices PXA27x and later processors support overlay1 and overlay2 on-top of the base framebuffer (although under-neath the base is also possible). They support palette and no-palette RGB formats, as well as YUV formats (only available on overlay2). These overlays have dedicated DMA channels and behave in a similar way as a framebuffer. This heavily simplified and re-structured work is based on the original pxafb_overlay.c (which is pending for mainline merge for a long time). The major problems with this pxafb_overlay.c are (if you are interested in the history): 1. heavily redundant (the control logics for overlay1 and overlay2 are actually identical except for some small operations, which are now abstracted into a 'pxafb_layer_ops' structure) 2. a lot of useless and un-tested code (two workarounds which are now fixed on mature silicons) 3. cursorfb is actually useless, hardware cursor should not be used this way, and the code was actually un-tested for a long time. The code in this patch should be self-explanatory, I tried to add minimum comments. As said, this is basically simplified, there are several things still on the pending list: 1. palette mode is un-supported and un-tested (although re-using the palette code of the base framebuffer is actually very easy now with previous clean-up patches) 2. fb_pan_display for overlay(s) is un-supported 3. the base framebuffer can actually be abstracted by 'pxafb_layer' as well, which will help further re-use of the code and keep a better and consistent structure. (This is the reason I named it 'pxafb_layer' instead of 'pxafb_overlay' or something alike) See Documentation/fb/pxafb.txt for additional usage information. Signed-off-by: Eric Miao <eric.miao@marvell.com> Cc: Rodolfo Giometti <giometti@linux.it> Signed-off-by: Eric Miao <ycmiao@ycmiao-hp520.(none)>
2008-12-23 17:49:43 +08:00
pxafb_overlay_exit(fbi);
unregister_framebuffer(info);
pxafb_disable_controller(fbi);
if (fbi->fb.cmap.len)
fb_dealloc_cmap(&fbi->fb.cmap);
irq = platform_get_irq(dev, 0);
free_irq(irq, fbi);
free_pages_exact(fbi->video_mem, fbi->video_mem_size);
dma_free_writecombine(&dev->dev, fbi->dma_buff_size,
fbi->dma_buff, fbi->dma_buff_phys);
iounmap(fbi->mmio_base);
r = platform_get_resource(dev, IORESOURCE_MEM, 0);
release_mem_region(r->start, resource_size(r));
clk_put(fbi->clk);
kfree(fbi);
return 0;
}
static struct platform_driver pxafb_driver = {
.probe = pxafb_probe,
.remove = __devexit_p(pxafb_remove),
.driver = {
.owner = THIS_MODULE,
.name = "pxa2xx-fb",
#ifdef CONFIG_PM
.pm = &pxafb_pm_ops,
#endif
},
};
static int __init pxafb_init(void)
{
if (pxafb_setup_options())
return -EINVAL;
return platform_driver_register(&pxafb_driver);
}
static void __exit pxafb_exit(void)
{
platform_driver_unregister(&pxafb_driver);
}
module_init(pxafb_init);
module_exit(pxafb_exit);
MODULE_DESCRIPTION("loadable framebuffer driver for PXA");
MODULE_LICENSE("GPL");