2
0
mirror of https://github.com/edk2-porting/linux-next.git synced 2024-12-27 14:43:58 +08:00
linux-next/drivers/video/fbmem.c

1624 lines
39 KiB
C
Raw Normal View History

/*
* linux/drivers/video/fbmem.c
*
* Copyright (C) 1994 Martin Schaller
*
* 2001 - Documented with DocBook
* - Brad Douglas <brad@neruo.com>
*
* 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.
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/compat.h>
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/smp_lock.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/mman.h>
#include <linux/tty.h>
#include <linux/init.h>
#include <linux/linux_logo.h>
#include <linux/proc_fs.h>
#include <linux/console.h>
#ifdef CONFIG_KMOD
#include <linux/kmod.h>
#endif
#include <linux/devfs_fs_kernel.h>
#include <linux/err.h>
#include <linux/device.h>
#include <linux/efi.h>
#if defined(__mc68000__) || defined(CONFIG_APUS)
#include <asm/setup.h>
#endif
#include <asm/io.h>
#include <asm/uaccess.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <linux/fb.h>
/*
* Frame buffer device initialization and setup routines
*/
#define FBPIXMAPSIZE (1024 * 8)
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
static BLOCKING_NOTIFIER_HEAD(fb_notifier_list);
struct fb_info *registered_fb[FB_MAX];
int num_registered_fb;
/*
* Helpers
*/
int fb_get_color_depth(struct fb_var_screeninfo *var,
struct fb_fix_screeninfo *fix)
{
int depth = 0;
if (fix->visual == FB_VISUAL_MONO01 ||
fix->visual == FB_VISUAL_MONO10)
depth = 1;
else {
if (var->green.length == var->blue.length &&
var->green.length == var->red.length &&
var->green.offset == var->blue.offset &&
var->green.offset == var->red.offset)
depth = var->green.length;
else
depth = var->green.length + var->red.length +
var->blue.length;
}
return depth;
}
EXPORT_SYMBOL(fb_get_color_depth);
/*
* Data padding functions.
*/
void fb_pad_aligned_buffer(u8 *dst, u32 d_pitch, u8 *src, u32 s_pitch, u32 height)
{
__fb_pad_aligned_buffer(dst, d_pitch, src, s_pitch, height);
}
EXPORT_SYMBOL(fb_pad_aligned_buffer);
void fb_pad_unaligned_buffer(u8 *dst, u32 d_pitch, u8 *src, u32 idx, u32 height,
u32 shift_high, u32 shift_low, u32 mod)
{
u8 mask = (u8) (0xfff << shift_high), tmp;
int i, j;
for (i = height; i--; ) {
for (j = 0; j < idx; j++) {
tmp = dst[j];
tmp &= mask;
tmp |= *src >> shift_low;
dst[j] = tmp;
tmp = *src << shift_high;
dst[j+1] = tmp;
src++;
}
tmp = dst[idx];
tmp &= mask;
tmp |= *src >> shift_low;
dst[idx] = tmp;
if (shift_high < mod) {
tmp = *src << shift_high;
dst[idx+1] = tmp;
}
src++;
dst += d_pitch;
}
}
EXPORT_SYMBOL(fb_pad_unaligned_buffer);
/*
* we need to lock this section since fb_cursor
* may use fb_imageblit()
*/
char* fb_get_buffer_offset(struct fb_info *info, struct fb_pixmap *buf, u32 size)
{
u32 align = buf->buf_align - 1, offset;
char *addr = buf->addr;
/* If IO mapped, we need to sync before access, no sharing of
* the pixmap is done
*/
if (buf->flags & FB_PIXMAP_IO) {
if (info->fbops->fb_sync && (buf->flags & FB_PIXMAP_SYNC))
info->fbops->fb_sync(info);
return addr;
}
/* See if we fit in the remaining pixmap space */
offset = buf->offset + align;
offset &= ~align;
if (offset + size > buf->size) {
/* We do not fit. In order to be able to re-use the buffer,
* we must ensure no asynchronous DMA'ing or whatever operation
* is in progress, we sync for that.
*/
if (info->fbops->fb_sync && (buf->flags & FB_PIXMAP_SYNC))
info->fbops->fb_sync(info);
offset = 0;
}
buf->offset = offset + size;
addr += offset;
return addr;
}
#ifdef CONFIG_LOGO
static inline unsigned safe_shift(unsigned d, int n)
{
return n < 0 ? d >> -n : d << n;
}
static void fb_set_logocmap(struct fb_info *info,
const struct linux_logo *logo)
{
struct fb_cmap palette_cmap;
u16 palette_green[16];
u16 palette_blue[16];
u16 palette_red[16];
int i, j, n;
const unsigned char *clut = logo->clut;
palette_cmap.start = 0;
palette_cmap.len = 16;
palette_cmap.red = palette_red;
palette_cmap.green = palette_green;
palette_cmap.blue = palette_blue;
palette_cmap.transp = NULL;
for (i = 0; i < logo->clutsize; i += n) {
n = logo->clutsize - i;
/* palette_cmap provides space for only 16 colors at once */
if (n > 16)
n = 16;
palette_cmap.start = 32 + i;
palette_cmap.len = n;
for (j = 0; j < n; ++j) {
palette_cmap.red[j] = clut[0] << 8 | clut[0];
palette_cmap.green[j] = clut[1] << 8 | clut[1];
palette_cmap.blue[j] = clut[2] << 8 | clut[2];
clut += 3;
}
fb_set_cmap(&palette_cmap, info);
}
}
static void fb_set_logo_truepalette(struct fb_info *info,
const struct linux_logo *logo,
u32 *palette)
{
unsigned char mask[9] = { 0,0x80,0xc0,0xe0,0xf0,0xf8,0xfc,0xfe,0xff };
unsigned char redmask, greenmask, bluemask;
int redshift, greenshift, blueshift;
int i;
const unsigned char *clut = logo->clut;
/*
* We have to create a temporary palette since console palette is only
* 16 colors long.
*/
/* Bug: Doesn't obey msb_right ... (who needs that?) */
redmask = mask[info->var.red.length < 8 ? info->var.red.length : 8];
greenmask = mask[info->var.green.length < 8 ? info->var.green.length : 8];
bluemask = mask[info->var.blue.length < 8 ? info->var.blue.length : 8];
redshift = info->var.red.offset - (8 - info->var.red.length);
greenshift = info->var.green.offset - (8 - info->var.green.length);
blueshift = info->var.blue.offset - (8 - info->var.blue.length);
for ( i = 0; i < logo->clutsize; i++) {
palette[i+32] = (safe_shift((clut[0] & redmask), redshift) |
safe_shift((clut[1] & greenmask), greenshift) |
safe_shift((clut[2] & bluemask), blueshift));
clut += 3;
}
}
static void fb_set_logo_directpalette(struct fb_info *info,
const struct linux_logo *logo,
u32 *palette)
{
int redshift, greenshift, blueshift;
int i;
redshift = info->var.red.offset;
greenshift = info->var.green.offset;
blueshift = info->var.blue.offset;
for (i = 32; i < logo->clutsize; i++)
palette[i] = i << redshift | i << greenshift | i << blueshift;
}
static void fb_set_logo(struct fb_info *info,
const struct linux_logo *logo, u8 *dst,
int depth)
{
int i, j, k;
const u8 *src = logo->data;
u8 xor = (info->fix.visual == FB_VISUAL_MONO01) ? 0xff : 0;
u8 fg = 1, d;
if (fb_get_color_depth(&info->var, &info->fix) == 3)
fg = 7;
if (info->fix.visual == FB_VISUAL_MONO01 ||
info->fix.visual == FB_VISUAL_MONO10)
fg = ~((u8) (0xfff << info->var.green.length));
switch (depth) {
case 4:
for (i = 0; i < logo->height; i++)
for (j = 0; j < logo->width; src++) {
*dst++ = *src >> 4;
j++;
if (j < logo->width) {
*dst++ = *src & 0x0f;
j++;
}
}
break;
case 1:
for (i = 0; i < logo->height; i++) {
for (j = 0; j < logo->width; src++) {
d = *src ^ xor;
for (k = 7; k >= 0; k--) {
*dst++ = ((d >> k) & 1) ? fg : 0;
j++;
}
}
}
break;
}
}
/*
* Three (3) kinds of logo maps exist. linux_logo_clut224 (>16 colors),
* linux_logo_vga16 (16 colors) and linux_logo_mono (2 colors). Depending on
* the visual format and color depth of the framebuffer, the DAC, the
* pseudo_palette, and the logo data will be adjusted accordingly.
*
* Case 1 - linux_logo_clut224:
* Color exceeds the number of console colors (16), thus we set the hardware DAC
* using fb_set_cmap() appropriately. The "needs_cmapreset" flag will be set.
*
* For visuals that require color info from the pseudo_palette, we also construct
* one for temporary use. The "needs_directpalette" or "needs_truepalette" flags
* will be set.
*
* Case 2 - linux_logo_vga16:
* The number of colors just matches the console colors, thus there is no need
* to set the DAC or the pseudo_palette. However, the bitmap is packed, ie,
* each byte contains color information for two pixels (upper and lower nibble).
* To be consistent with fb_imageblit() usage, we therefore separate the two
* nibbles into separate bytes. The "depth" flag will be set to 4.
*
* Case 3 - linux_logo_mono:
* This is similar with Case 2. Each byte contains information for 8 pixels.
* We isolate each bit and expand each into a byte. The "depth" flag will
* be set to 1.
*/
static struct logo_data {
int depth;
int needs_directpalette;
int needs_truepalette;
int needs_cmapreset;
const struct linux_logo *logo;
} fb_logo;
static void fb_rotate_logo_ud(const u8 *in, u8 *out, u32 width, u32 height)
{
u32 size = width * height, i;
out += size - 1;
for (i = size; i--; )
*out-- = *in++;
}
static void fb_rotate_logo_cw(const u8 *in, u8 *out, u32 width, u32 height)
{
int i, j, h = height - 1;
for (i = 0; i < height; i++)
for (j = 0; j < width; j++)
out[height * j + h - i] = *in++;
}
static void fb_rotate_logo_ccw(const u8 *in, u8 *out, u32 width, u32 height)
{
int i, j, w = width - 1;
for (i = 0; i < height; i++)
for (j = 0; j < width; j++)
out[height * (w - j) + i] = *in++;
}
static void fb_rotate_logo(struct fb_info *info, u8 *dst,
struct fb_image *image, int rotate)
{
u32 tmp;
if (rotate == FB_ROTATE_UD) {
fb_rotate_logo_ud(image->data, dst, image->width,
image->height);
image->dx = info->var.xres - image->width;
image->dy = info->var.yres - image->height;
} else if (rotate == FB_ROTATE_CW) {
fb_rotate_logo_cw(image->data, dst, image->width,
image->height);
tmp = image->width;
image->width = image->height;
image->height = tmp;
image->dx = info->var.xres - image->width;
} else if (rotate == FB_ROTATE_CCW) {
fb_rotate_logo_ccw(image->data, dst, image->width,
image->height);
tmp = image->width;
image->width = image->height;
image->height = tmp;
image->dy = info->var.yres - image->height;
}
image->data = dst;
}
static void fb_do_show_logo(struct fb_info *info, struct fb_image *image,
int rotate)
{
int x;
if (rotate == FB_ROTATE_UR) {
for (x = 0; x < num_online_cpus() &&
x * (fb_logo.logo->width + 8) <=
info->var.xres - fb_logo.logo->width; x++) {
info->fbops->fb_imageblit(info, image);
image->dx += fb_logo.logo->width + 8;
}
} else if (rotate == FB_ROTATE_UD) {
for (x = 0; x < num_online_cpus() &&
x * (fb_logo.logo->width + 8) <=
info->var.xres - fb_logo.logo->width; x++) {
info->fbops->fb_imageblit(info, image);
image->dx -= fb_logo.logo->width + 8;
}
} else if (rotate == FB_ROTATE_CW) {
for (x = 0; x < num_online_cpus() &&
x * (fb_logo.logo->width + 8) <=
info->var.yres - fb_logo.logo->width; x++) {
info->fbops->fb_imageblit(info, image);
image->dy += fb_logo.logo->width + 8;
}
} else if (rotate == FB_ROTATE_CCW) {
for (x = 0; x < num_online_cpus() &&
x * (fb_logo.logo->width + 8) <=
info->var.yres - fb_logo.logo->width; x++) {
info->fbops->fb_imageblit(info, image);
image->dy -= fb_logo.logo->width + 8;
}
}
}
int fb_prepare_logo(struct fb_info *info, int rotate)
{
int depth = fb_get_color_depth(&info->var, &info->fix);
int yres;
memset(&fb_logo, 0, sizeof(struct logo_data));
if (info->flags & FBINFO_MISC_TILEBLITTING)
return 0;
if (info->fix.visual == FB_VISUAL_DIRECTCOLOR) {
depth = info->var.blue.length;
if (info->var.red.length < depth)
depth = info->var.red.length;
if (info->var.green.length < depth)
depth = info->var.green.length;
}
if (info->fix.visual == FB_VISUAL_STATIC_PSEUDOCOLOR && depth > 4) {
/* assume console colormap */
depth = 4;
}
if (depth >= 8) {
switch (info->fix.visual) {
case FB_VISUAL_TRUECOLOR:
fb_logo.needs_truepalette = 1;
break;
case FB_VISUAL_DIRECTCOLOR:
fb_logo.needs_directpalette = 1;
fb_logo.needs_cmapreset = 1;
break;
case FB_VISUAL_PSEUDOCOLOR:
fb_logo.needs_cmapreset = 1;
break;
}
}
/* Return if no suitable logo was found */
fb_logo.logo = fb_find_logo(depth);
if (!fb_logo.logo) {
return 0;
}
if (rotate == FB_ROTATE_UR || rotate == FB_ROTATE_UD)
yres = info->var.yres;
else
yres = info->var.xres;
if (fb_logo.logo->height > yres) {
fb_logo.logo = NULL;
return 0;
}
/* What depth we asked for might be different from what we get */
if (fb_logo.logo->type == LINUX_LOGO_CLUT224)
fb_logo.depth = 8;
else if (fb_logo.logo->type == LINUX_LOGO_VGA16)
fb_logo.depth = 4;
else
fb_logo.depth = 1;
return fb_logo.logo->height;
}
int fb_show_logo(struct fb_info *info, int rotate)
{
u32 *palette = NULL, *saved_pseudo_palette = NULL;
unsigned char *logo_new = NULL, *logo_rotate = NULL;
struct fb_image image;
/* Return if the frame buffer is not mapped or suspended */
if (fb_logo.logo == NULL || info->state != FBINFO_STATE_RUNNING)
return 0;
image.depth = 8;
image.data = fb_logo.logo->data;
if (fb_logo.needs_cmapreset)
fb_set_logocmap(info, fb_logo.logo);
if (fb_logo.needs_truepalette ||
fb_logo.needs_directpalette) {
palette = kmalloc(256 * 4, GFP_KERNEL);
if (palette == NULL)
return 0;
if (fb_logo.needs_truepalette)
fb_set_logo_truepalette(info, fb_logo.logo, palette);
else
fb_set_logo_directpalette(info, fb_logo.logo, palette);
saved_pseudo_palette = info->pseudo_palette;
info->pseudo_palette = palette;
}
if (fb_logo.depth <= 4) {
logo_new = kmalloc(fb_logo.logo->width * fb_logo.logo->height,
GFP_KERNEL);
if (logo_new == NULL) {
kfree(palette);
if (saved_pseudo_palette)
info->pseudo_palette = saved_pseudo_palette;
return 0;
}
image.data = logo_new;
fb_set_logo(info, fb_logo.logo, logo_new, fb_logo.depth);
}
image.dx = 0;
image.dy = 0;
image.width = fb_logo.logo->width;
image.height = fb_logo.logo->height;
if (rotate) {
logo_rotate = kmalloc(fb_logo.logo->width *
fb_logo.logo->height, GFP_KERNEL);
if (logo_rotate)
fb_rotate_logo(info, logo_rotate, &image, rotate);
}
fb_do_show_logo(info, &image, rotate);
kfree(palette);
if (saved_pseudo_palette != NULL)
info->pseudo_palette = saved_pseudo_palette;
kfree(logo_new);
kfree(logo_rotate);
return fb_logo.logo->height;
}
#else
int fb_prepare_logo(struct fb_info *info, int rotate) { return 0; }
int fb_show_logo(struct fb_info *info, int rotate) { return 0; }
#endif /* CONFIG_LOGO */
static int fbmem_read_proc(char *buf, char **start, off_t offset,
int len, int *eof, void *private)
{
struct fb_info **fi;
int clen;
clen = 0;
for (fi = registered_fb; fi < &registered_fb[FB_MAX] && len < 4000; fi++)
if (*fi)
clen += sprintf(buf + clen, "%d %s\n",
(*fi)->node,
(*fi)->fix.id);
*start = buf + offset;
if (clen > offset)
clen -= offset;
else
clen = 0;
return clen < len ? clen : len;
}
static ssize_t
fb_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
{
unsigned long p = *ppos;
struct inode *inode = file->f_dentry->d_inode;
int fbidx = iminor(inode);
struct fb_info *info = registered_fb[fbidx];
u32 *buffer, *dst;
u32 __iomem *src;
int c, i, cnt = 0, err = 0;
unsigned long total_size;
if (!info || ! info->screen_base)
return -ENODEV;
if (info->state != FBINFO_STATE_RUNNING)
return -EPERM;
if (info->fbops->fb_read)
return info->fbops->fb_read(file, buf, count, ppos);
total_size = info->screen_size;
if (total_size == 0)
total_size = info->fix.smem_len;
if (p >= total_size)
return 0;
if (count >= total_size)
count = total_size;
if (count + p > total_size)
count = total_size - p;
buffer = kmalloc((count > PAGE_SIZE) ? PAGE_SIZE : count,
GFP_KERNEL);
if (!buffer)
return -ENOMEM;
src = (u32 __iomem *) (info->screen_base + p);
if (info->fbops->fb_sync)
info->fbops->fb_sync(info);
while (count) {
c = (count > PAGE_SIZE) ? PAGE_SIZE : count;
dst = buffer;
for (i = c >> 2; i--; )
*dst++ = fb_readl(src++);
if (c & 3) {
u8 *dst8 = (u8 *) dst;
u8 __iomem *src8 = (u8 __iomem *) src;
for (i = c & 3; i--;)
*dst8++ = fb_readb(src8++);
src = (u32 __iomem *) src8;
}
if (copy_to_user(buf, buffer, c)) {
err = -EFAULT;
break;
}
*ppos += c;
buf += c;
cnt += c;
count -= c;
}
kfree(buffer);
return (err) ? err : cnt;
}
static ssize_t
fb_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
{
unsigned long p = *ppos;
struct inode *inode = file->f_dentry->d_inode;
int fbidx = iminor(inode);
struct fb_info *info = registered_fb[fbidx];
u32 *buffer, *src;
u32 __iomem *dst;
int c, i, cnt = 0, err = 0;
unsigned long total_size;
if (!info || !info->screen_base)
return -ENODEV;
if (info->state != FBINFO_STATE_RUNNING)
return -EPERM;
if (info->fbops->fb_write)
return info->fbops->fb_write(file, buf, count, ppos);
total_size = info->screen_size;
if (total_size == 0)
total_size = info->fix.smem_len;
if (p > total_size)
return -EFBIG;
if (count > total_size) {
err = -EFBIG;
count = total_size;
}
if (count + p > total_size) {
if (!err)
err = -ENOSPC;
count = total_size - p;
}
buffer = kmalloc((count > PAGE_SIZE) ? PAGE_SIZE : count,
GFP_KERNEL);
if (!buffer)
return -ENOMEM;
dst = (u32 __iomem *) (info->screen_base + p);
if (info->fbops->fb_sync)
info->fbops->fb_sync(info);
while (count) {
c = (count > PAGE_SIZE) ? PAGE_SIZE : count;
src = buffer;
if (copy_from_user(src, buf, c)) {
err = -EFAULT;
break;
}
for (i = c >> 2; i--; )
fb_writel(*src++, dst++);
if (c & 3) {
u8 *src8 = (u8 *) src;
u8 __iomem *dst8 = (u8 __iomem *) dst;
for (i = c & 3; i--; )
fb_writeb(*src8++, dst8++);
dst = (u32 __iomem *) dst8;
}
*ppos += c;
buf += c;
cnt += c;
count -= c;
}
kfree(buffer);
return (cnt) ? cnt : err;
}
#ifdef CONFIG_KMOD
static void try_to_load(int fb)
{
request_module("fb%d", fb);
}
#endif /* CONFIG_KMOD */
int
fb_pan_display(struct fb_info *info, struct fb_var_screeninfo *var)
{
struct fb_fix_screeninfo *fix = &info->fix;
int xoffset = var->xoffset;
int yoffset = var->yoffset;
int err = 0, yres = info->var.yres;
if (var->yoffset > 0) {
if (var->vmode & FB_VMODE_YWRAP) {
if (!fix->ywrapstep || (var->yoffset % fix->ywrapstep))
err = -EINVAL;
else
yres = 0;
} else if (!fix->ypanstep || (var->yoffset % fix->ypanstep))
err = -EINVAL;
}
if (var->xoffset > 0 && (!fix->xpanstep ||
(var->xoffset % fix->xpanstep)))
err = -EINVAL;
if (err || !info->fbops->fb_pan_display || xoffset < 0 ||
yoffset < 0 || var->yoffset + yres > info->var.yres_virtual ||
var->xoffset + info->var.xres > info->var.xres_virtual)
return -EINVAL;
if ((err = info->fbops->fb_pan_display(var, info)))
return err;
info->var.xoffset = var->xoffset;
info->var.yoffset = var->yoffset;
if (var->vmode & FB_VMODE_YWRAP)
info->var.vmode |= FB_VMODE_YWRAP;
else
info->var.vmode &= ~FB_VMODE_YWRAP;
return 0;
}
int
fb_set_var(struct fb_info *info, struct fb_var_screeninfo *var)
{
int err, flags = info->flags;
if (var->activate & FB_ACTIVATE_INV_MODE) {
struct fb_videomode mode1, mode2;
int ret = 0;
fb_var_to_videomode(&mode1, var);
fb_var_to_videomode(&mode2, &info->var);
/* make sure we don't delete the videomode of current var */
ret = fb_mode_is_equal(&mode1, &mode2);
if (!ret) {
struct fb_event event;
event.info = info;
event.data = &mode1;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
ret = blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_MODE_DELETE, &event);
}
if (!ret)
fb_delete_videomode(&mode1, &info->modelist);
return ret;
}
if ((var->activate & FB_ACTIVATE_FORCE) ||
memcmp(&info->var, var, sizeof(struct fb_var_screeninfo))) {
if (!info->fbops->fb_check_var) {
*var = info->var;
return 0;
}
if ((err = info->fbops->fb_check_var(var, info)))
return err;
if ((var->activate & FB_ACTIVATE_MASK) == FB_ACTIVATE_NOW) {
struct fb_videomode mode;
int err = 0;
info->var = *var;
if (info->fbops->fb_set_par)
info->fbops->fb_set_par(info);
fb_pan_display(info, &info->var);
fb_set_cmap(&info->cmap, info);
fb_var_to_videomode(&mode, &info->var);
if (info->modelist.prev && info->modelist.next &&
!list_empty(&info->modelist))
err = fb_add_videomode(&mode, &info->modelist);
if (!err && (flags & FBINFO_MISC_USEREVENT)) {
struct fb_event event;
int evnt = (var->activate & FB_ACTIVATE_ALL) ?
FB_EVENT_MODE_CHANGE_ALL :
FB_EVENT_MODE_CHANGE;
info->flags &= ~FBINFO_MISC_USEREVENT;
event.info = info;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
evnt, &event);
}
}
}
return 0;
}
int
fb_blank(struct fb_info *info, int blank)
{
int ret = -EINVAL;
if (blank > FB_BLANK_POWERDOWN)
blank = FB_BLANK_POWERDOWN;
if (info->fbops->fb_blank)
ret = info->fbops->fb_blank(blank, info);
if (!ret) {
struct fb_event event;
event.info = info;
event.data = &blank;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_BLANK, &event);
}
return ret;
}
static int
fb_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
unsigned long arg)
{
int fbidx = iminor(inode);
struct fb_info *info = registered_fb[fbidx];
struct fb_ops *fb = info->fbops;
struct fb_var_screeninfo var;
struct fb_fix_screeninfo fix;
struct fb_con2fbmap con2fb;
struct fb_cmap_user cmap;
struct fb_event event;
void __user *argp = (void __user *)arg;
int i;
if (!fb)
return -ENODEV;
switch (cmd) {
case FBIOGET_VSCREENINFO:
return copy_to_user(argp, &info->var,
sizeof(var)) ? -EFAULT : 0;
case FBIOPUT_VSCREENINFO:
if (copy_from_user(&var, argp, sizeof(var)))
return -EFAULT;
acquire_console_sem();
info->flags |= FBINFO_MISC_USEREVENT;
i = fb_set_var(info, &var);
info->flags &= ~FBINFO_MISC_USEREVENT;
release_console_sem();
if (i) return i;
if (copy_to_user(argp, &var, sizeof(var)))
return -EFAULT;
return 0;
case FBIOGET_FSCREENINFO:
return copy_to_user(argp, &info->fix,
sizeof(fix)) ? -EFAULT : 0;
case FBIOPUTCMAP:
if (copy_from_user(&cmap, argp, sizeof(cmap)))
return -EFAULT;
return (fb_set_user_cmap(&cmap, info));
case FBIOGETCMAP:
if (copy_from_user(&cmap, argp, sizeof(cmap)))
return -EFAULT;
return fb_cmap_to_user(&info->cmap, &cmap);
case FBIOPAN_DISPLAY:
if (copy_from_user(&var, argp, sizeof(var)))
return -EFAULT;
acquire_console_sem();
i = fb_pan_display(info, &var);
release_console_sem();
if (i)
return i;
if (copy_to_user(argp, &var, sizeof(var)))
return -EFAULT;
return 0;
case FBIO_CURSOR:
return -EINVAL;
case FBIOGET_CON2FBMAP:
if (copy_from_user(&con2fb, argp, sizeof(con2fb)))
return -EFAULT;
if (con2fb.console < 1 || con2fb.console > MAX_NR_CONSOLES)
return -EINVAL;
con2fb.framebuffer = -1;
event.info = info;
event.data = &con2fb;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_GET_CONSOLE_MAP, &event);
return copy_to_user(argp, &con2fb,
sizeof(con2fb)) ? -EFAULT : 0;
case FBIOPUT_CON2FBMAP:
if (copy_from_user(&con2fb, argp, sizeof(con2fb)))
return - EFAULT;
if (con2fb.console < 0 || con2fb.console > MAX_NR_CONSOLES)
return -EINVAL;
if (con2fb.framebuffer < 0 || con2fb.framebuffer >= FB_MAX)
return -EINVAL;
#ifdef CONFIG_KMOD
if (!registered_fb[con2fb.framebuffer])
try_to_load(con2fb.framebuffer);
#endif /* CONFIG_KMOD */
if (!registered_fb[con2fb.framebuffer])
return -EINVAL;
event.info = info;
event.data = &con2fb;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
return blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_SET_CONSOLE_MAP,
&event);
case FBIOBLANK:
acquire_console_sem();
info->flags |= FBINFO_MISC_USEREVENT;
i = fb_blank(info, arg);
info->flags &= ~FBINFO_MISC_USEREVENT;
release_console_sem();
return i;
default:
if (fb->fb_ioctl == NULL)
return -EINVAL;
return fb->fb_ioctl(info, cmd, arg);
}
}
#ifdef CONFIG_COMPAT
struct fb_fix_screeninfo32 {
char id[16];
compat_caddr_t smem_start;
u32 smem_len;
u32 type;
u32 type_aux;
u32 visual;
u16 xpanstep;
u16 ypanstep;
u16 ywrapstep;
u32 line_length;
compat_caddr_t mmio_start;
u32 mmio_len;
u32 accel;
u16 reserved[3];
};
struct fb_cmap32 {
u32 start;
u32 len;
compat_caddr_t red;
compat_caddr_t green;
compat_caddr_t blue;
compat_caddr_t transp;
};
static int fb_getput_cmap(struct inode *inode, struct file *file,
unsigned int cmd, unsigned long arg)
{
struct fb_cmap_user __user *cmap;
struct fb_cmap32 __user *cmap32;
__u32 data;
int err;
cmap = compat_alloc_user_space(sizeof(*cmap));
cmap32 = compat_ptr(arg);
if (copy_in_user(&cmap->start, &cmap32->start, 2 * sizeof(__u32)))
return -EFAULT;
if (get_user(data, &cmap32->red) ||
put_user(compat_ptr(data), &cmap->red) ||
get_user(data, &cmap32->green) ||
put_user(compat_ptr(data), &cmap->green) ||
get_user(data, &cmap32->blue) ||
put_user(compat_ptr(data), &cmap->blue) ||
get_user(data, &cmap32->transp) ||
put_user(compat_ptr(data), &cmap->transp))
return -EFAULT;
err = fb_ioctl(inode, file, cmd, (unsigned long) cmap);
if (!err) {
if (copy_in_user(&cmap32->start,
&cmap->start,
2 * sizeof(__u32)))
err = -EFAULT;
}
return err;
}
static int do_fscreeninfo_to_user(struct fb_fix_screeninfo *fix,
struct fb_fix_screeninfo32 __user *fix32)
{
__u32 data;
int err;
err = copy_to_user(&fix32->id, &fix->id, sizeof(fix32->id));
data = (__u32) (unsigned long) fix->smem_start;
err |= put_user(data, &fix32->smem_start);
err |= put_user(fix->smem_len, &fix32->smem_len);
err |= put_user(fix->type, &fix32->type);
err |= put_user(fix->type_aux, &fix32->type_aux);
err |= put_user(fix->visual, &fix32->visual);
err |= put_user(fix->xpanstep, &fix32->xpanstep);
err |= put_user(fix->ypanstep, &fix32->ypanstep);
err |= put_user(fix->ywrapstep, &fix32->ywrapstep);
err |= put_user(fix->line_length, &fix32->line_length);
data = (__u32) (unsigned long) fix->mmio_start;
err |= put_user(data, &fix32->mmio_start);
err |= put_user(fix->mmio_len, &fix32->mmio_len);
err |= put_user(fix->accel, &fix32->accel);
err |= copy_to_user(fix32->reserved, fix->reserved,
sizeof(fix->reserved));
return err;
}
static int fb_get_fscreeninfo(struct inode *inode, struct file *file,
unsigned int cmd, unsigned long arg)
{
mm_segment_t old_fs;
struct fb_fix_screeninfo fix;
struct fb_fix_screeninfo32 __user *fix32;
int err;
fix32 = compat_ptr(arg);
old_fs = get_fs();
set_fs(KERNEL_DS);
err = fb_ioctl(inode, file, cmd, (unsigned long) &fix);
set_fs(old_fs);
if (!err)
err = do_fscreeninfo_to_user(&fix, fix32);
return err;
}
static long
fb_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
struct inode *inode = file->f_dentry->d_inode;
int fbidx = iminor(inode);
struct fb_info *info = registered_fb[fbidx];
struct fb_ops *fb = info->fbops;
long ret = -ENOIOCTLCMD;
lock_kernel();
switch(cmd) {
case FBIOGET_VSCREENINFO:
case FBIOPUT_VSCREENINFO:
case FBIOPAN_DISPLAY:
case FBIOGET_CON2FBMAP:
case FBIOPUT_CON2FBMAP:
arg = (unsigned long) compat_ptr(arg);
case FBIOBLANK:
ret = fb_ioctl(inode, file, cmd, arg);
break;
case FBIOGET_FSCREENINFO:
ret = fb_get_fscreeninfo(inode, file, cmd, arg);
break;
case FBIOGETCMAP:
case FBIOPUTCMAP:
ret = fb_getput_cmap(inode, file, cmd, arg);
break;
default:
if (fb->fb_compat_ioctl)
ret = fb->fb_compat_ioctl(info, cmd, arg);
break;
}
unlock_kernel();
return ret;
}
#endif
static int
fb_mmap(struct file *file, struct vm_area_struct * vma)
{
int fbidx = iminor(file->f_dentry->d_inode);
struct fb_info *info = registered_fb[fbidx];
struct fb_ops *fb = info->fbops;
unsigned long off;
#if !defined(__sparc__) || defined(__sparc_v9__)
unsigned long start;
u32 len;
#endif
if (vma->vm_pgoff > (~0UL >> PAGE_SHIFT))
return -EINVAL;
off = vma->vm_pgoff << PAGE_SHIFT;
if (!fb)
return -ENODEV;
if (fb->fb_mmap) {
int res;
lock_kernel();
res = fb->fb_mmap(info, vma);
unlock_kernel();
return res;
}
#if defined(__sparc__) && !defined(__sparc_v9__)
/* Should never get here, all fb drivers should have their own
mmap routines */
return -EINVAL;
#else
/* !sparc32... */
lock_kernel();
/* frame buffer memory */
start = info->fix.smem_start;
len = PAGE_ALIGN((start & ~PAGE_MASK) + info->fix.smem_len);
if (off >= len) {
/* memory mapped io */
off -= len;
if (info->var.accel_flags) {
unlock_kernel();
return -EINVAL;
}
start = info->fix.mmio_start;
len = PAGE_ALIGN((start & ~PAGE_MASK) + info->fix.mmio_len);
}
unlock_kernel();
start &= PAGE_MASK;
if ((vma->vm_end - vma->vm_start + off) > len)
return -EINVAL;
off += start;
vma->vm_pgoff = off >> PAGE_SHIFT;
/* This is an IO map - tell maydump to skip this VMA */
vma->vm_flags |= VM_IO | VM_RESERVED;
#if defined(__mc68000__)
#if defined(CONFIG_SUN3)
pgprot_val(vma->vm_page_prot) |= SUN3_PAGE_NOCACHE;
#elif defined(CONFIG_MMU)
if (CPU_IS_020_OR_030)
pgprot_val(vma->vm_page_prot) |= _PAGE_NOCACHE030;
if (CPU_IS_040_OR_060) {
pgprot_val(vma->vm_page_prot) &= _CACHEMASK040;
/* Use no-cache mode, serialized */
pgprot_val(vma->vm_page_prot) |= _PAGE_NOCACHE_S;
}
#endif
#elif defined(__powerpc__)
vma->vm_page_prot = phys_mem_access_prot(file, off >> PAGE_SHIFT,
vma->vm_end - vma->vm_start,
vma->vm_page_prot);
#elif defined(__alpha__)
/* Caching is off in the I/O space quadrant by design. */
#elif defined(__i386__) || defined(__x86_64__)
if (boot_cpu_data.x86 > 3)
pgprot_val(vma->vm_page_prot) |= _PAGE_PCD;
#elif defined(__mips__) || defined(__sparc_v9__)
vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
#elif defined(__hppa__)
pgprot_val(vma->vm_page_prot) |= _PAGE_NO_CACHE;
#elif defined(__arm__) || defined(__sh__) || defined(__m32r__)
vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
#elif defined(__ia64__)
if (efi_range_is_wc(vma->vm_start, vma->vm_end - vma->vm_start))
vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
else
vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
#else
#warning What do we have to do here??
#endif
if (io_remap_pfn_range(vma, vma->vm_start, off >> PAGE_SHIFT,
vma->vm_end - vma->vm_start, vma->vm_page_prot))
return -EAGAIN;
return 0;
#endif /* !sparc32 */
}
static int
fb_open(struct inode *inode, struct file *file)
{
int fbidx = iminor(inode);
struct fb_info *info;
int res = 0;
if (fbidx >= FB_MAX)
return -ENODEV;
#ifdef CONFIG_KMOD
if (!(info = registered_fb[fbidx]))
try_to_load(fbidx);
#endif /* CONFIG_KMOD */
if (!(info = registered_fb[fbidx]))
return -ENODEV;
if (!try_module_get(info->fbops->owner))
return -ENODEV;
file->private_data = info;
if (info->fbops->fb_open) {
res = info->fbops->fb_open(info,1);
if (res)
module_put(info->fbops->owner);
}
return res;
}
static int
fb_release(struct inode *inode, struct file *file)
{
struct fb_info * const info = file->private_data;
lock_kernel();
if (info->fbops->fb_release)
info->fbops->fb_release(info,1);
module_put(info->fbops->owner);
unlock_kernel();
return 0;
}
static struct file_operations fb_fops = {
.owner = THIS_MODULE,
.read = fb_read,
.write = fb_write,
.ioctl = fb_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = fb_compat_ioctl,
#endif
.mmap = fb_mmap,
.open = fb_open,
.release = fb_release,
#ifdef HAVE_ARCH_FB_UNMAPPED_AREA
.get_unmapped_area = get_fb_unmapped_area,
#endif
};
static struct class *fb_class;
/**
* register_framebuffer - registers a frame buffer device
* @fb_info: frame buffer info structure
*
* Registers a frame buffer device @fb_info.
*
* Returns negative errno on error, or zero for success.
*
*/
int
register_framebuffer(struct fb_info *fb_info)
{
int i;
struct fb_event event;
2005-09-10 04:09:58 +08:00
struct fb_videomode mode;
if (num_registered_fb == FB_MAX)
return -ENXIO;
num_registered_fb++;
for (i = 0 ; i < FB_MAX; i++)
if (!registered_fb[i])
break;
fb_info->node = i;
fb_info->class_device = class_device_create(fb_class, NULL, MKDEV(FB_MAJOR, i),
fb_info->device, "fb%d", i);
if (IS_ERR(fb_info->class_device)) {
/* Not fatal */
printk(KERN_WARNING "Unable to create class_device for framebuffer %d; errno = %ld\n", i, PTR_ERR(fb_info->class_device));
fb_info->class_device = NULL;
} else
fb_init_class_device(fb_info);
if (fb_info->pixmap.addr == NULL) {
fb_info->pixmap.addr = kmalloc(FBPIXMAPSIZE, GFP_KERNEL);
if (fb_info->pixmap.addr) {
fb_info->pixmap.size = FBPIXMAPSIZE;
fb_info->pixmap.buf_align = 1;
fb_info->pixmap.scan_align = 1;
fb_info->pixmap.access_align = 32;
fb_info->pixmap.flags = FB_PIXMAP_DEFAULT;
}
}
fb_info->pixmap.offset = 0;
2005-09-10 04:09:58 +08:00
if (!fb_info->modelist.prev || !fb_info->modelist.next)
INIT_LIST_HEAD(&fb_info->modelist);
2005-09-10 04:09:58 +08:00
fb_var_to_videomode(&mode, &fb_info->var);
fb_add_videomode(&mode, &fb_info->modelist);
registered_fb[i] = fb_info;
devfs_mk_cdev(MKDEV(FB_MAJOR, i),
S_IFCHR | S_IRUGO | S_IWUGO, "fb/%d", i);
event.info = fb_info;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_FB_REGISTERED, &event);
return 0;
}
/**
* unregister_framebuffer - releases a frame buffer device
* @fb_info: frame buffer info structure
*
* Unregisters a frame buffer device @fb_info.
*
* Returns negative errno on error, or zero for success.
*
*/
int
unregister_framebuffer(struct fb_info *fb_info)
{
int i;
i = fb_info->node;
if (!registered_fb[i])
return -EINVAL;
devfs_remove("fb/%d", i);
if (fb_info->pixmap.addr && (fb_info->pixmap.flags & FB_PIXMAP_DEFAULT))
kfree(fb_info->pixmap.addr);
fb_destroy_modelist(&fb_info->modelist);
registered_fb[i]=NULL;
num_registered_fb--;
fb_cleanup_class_device(fb_info);
class_device_destroy(fb_class, MKDEV(FB_MAJOR, i));
return 0;
}
/**
* fb_register_client - register a client notifier
* @nb: notifier block to callback on events
*/
int fb_register_client(struct notifier_block *nb)
{
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
return blocking_notifier_chain_register(&fb_notifier_list, nb);
}
/**
* fb_unregister_client - unregister a client notifier
* @nb: notifier block to callback on events
*/
int fb_unregister_client(struct notifier_block *nb)
{
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
return blocking_notifier_chain_unregister(&fb_notifier_list, nb);
}
/**
* fb_set_suspend - low level driver signals suspend
* @info: framebuffer affected
* @state: 0 = resuming, !=0 = suspending
*
* This is meant to be used by low level drivers to
* signal suspend/resume to the core & clients.
* It must be called with the console semaphore held
*/
void fb_set_suspend(struct fb_info *info, int state)
{
struct fb_event event;
event.info = info;
if (state) {
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_SUSPEND, &event);
info->state = FBINFO_STATE_SUSPENDED;
} else {
info->state = FBINFO_STATE_RUNNING;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_RESUME, &event);
}
}
/**
* fbmem_init - init frame buffer subsystem
*
* Initialize the frame buffer subsystem.
*
* NOTE: This function is _only_ to be called by drivers/char/mem.c.
*
*/
static int __init
fbmem_init(void)
{
create_proc_read_entry("fb", 0, NULL, fbmem_read_proc, NULL);
devfs_mk_dir("fb");
if (register_chrdev(FB_MAJOR,"fb",&fb_fops))
printk("unable to get major %d for fb devs\n", FB_MAJOR);
fb_class = class_create(THIS_MODULE, "graphics");
if (IS_ERR(fb_class)) {
printk(KERN_WARNING "Unable to create fb class; errno = %ld\n", PTR_ERR(fb_class));
fb_class = NULL;
}
return 0;
}
#ifdef MODULE
module_init(fbmem_init);
static void __exit
fbmem_exit(void)
{
class_destroy(fb_class);
unregister_chrdev(FB_MAJOR, "fb");
}
module_exit(fbmem_exit);
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("Framebuffer base");
#else
subsys_initcall(fbmem_init);
#endif
int fb_new_modelist(struct fb_info *info)
{
struct fb_event event;
struct fb_var_screeninfo var = info->var;
struct list_head *pos, *n;
struct fb_modelist *modelist;
struct fb_videomode *m, mode;
int err = 1;
list_for_each_safe(pos, n, &info->modelist) {
modelist = list_entry(pos, struct fb_modelist, list);
m = &modelist->mode;
fb_videomode_to_var(&var, m);
var.activate = FB_ACTIVATE_TEST;
err = fb_set_var(info, &var);
fb_var_to_videomode(&mode, &var);
if (err || !fb_mode_is_equal(m, &mode)) {
list_del(pos);
kfree(pos);
}
}
err = 1;
if (!list_empty(&info->modelist)) {
event.info = info;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
err = blocking_notifier_call_chain(&fb_notifier_list,
FB_EVENT_NEW_MODELIST,
&event);
}
return err;
}
/**
* fb_con_duit - user<->fbcon passthrough
* @info: struct fb_info
* @event: notification event to be passed to fbcon
* @data: private data
*
* DESCRIPTION
* This function is an fbcon-user event passing channel
* which bypasses fbdev. This is hopefully temporary
* until a user interface for fbcon is created
*/
int fb_con_duit(struct fb_info *info, int event, void *data)
{
struct fb_event evnt;
evnt.info = info;
evnt.data = data;
[PATCH] Notifier chain update: API changes The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-27 17:16:30 +08:00
return blocking_notifier_call_chain(&fb_notifier_list, event, &evnt);
}
static char *video_options[FB_MAX];
static int ofonly;
[PATCH] DocBook: changes and extensions to the kernel documentation I have recompiled Linux kernel 2.6.11.5 documentation for me and our university students again. The documentation could be extended for more sources which are equipped by structured comments for recent 2.6 kernels. I have tried to proceed with that task. I have done that more times from 2.6.0 time and it gets boring to do same changes again and again. Linux kernel compiles after changes for i386 and ARM targets. I have added references to some more files into kernel-api book, I have added some section names as well. So please, check that changes do not break something and that categories are not too much skewed. I have changed kernel-doc to accept "fastcall" and "asmlinkage" words reserved by kernel convention. Most of the other changes are modifications in the comments to make kernel-doc happy, accept some parameters description and do not bail out on errors. Changed <pid> to @pid in the description, moved some #ifdef before comments to correct function to comments bindings, etc. You can see result of the modified documentation build at http://cmp.felk.cvut.cz/~pisa/linux/lkdb-2.6.11.tar.gz Some more sources are ready to be included into kernel-doc generated documentation. Sources has been added into kernel-api for now. Some more section names added and probably some more chaos introduced as result of quick cleanup work. Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz> Signed-off-by: Martin Waitz <tali@admingilde.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-05-01 23:59:25 +08:00
extern const char *global_mode_option;
/**
* fb_get_options - get kernel boot parameters
* @name: framebuffer name as it would appear in
* the boot parameter line
* (video=<name>:<options>)
* @option: the option will be stored here
*
* NOTE: Needed to maintain backwards compatibility
*/
int fb_get_options(char *name, char **option)
{
char *opt, *options = NULL;
int opt_len, retval = 0;
int name_len = strlen(name), i;
if (name_len && ofonly && strncmp(name, "offb", 4))
retval = 1;
if (name_len && !retval) {
for (i = 0; i < FB_MAX; i++) {
if (video_options[i] == NULL)
continue;
opt_len = strlen(video_options[i]);
if (!opt_len)
continue;
opt = video_options[i];
if (!strncmp(name, opt, name_len) &&
opt[name_len] == ':')
options = opt + name_len + 1;
}
}
if (options && !strncmp(options, "off", 3))
retval = 1;
if (option)
*option = options;
return retval;
}
#ifndef MODULE
/**
* video_setup - process command line options
* @options: string of options
*
* Process command line options for frame buffer subsystem.
*
* NOTE: This function is a __setup and __init function.
* It only stores the options. Drivers have to call
* fb_get_options() as necessary.
*
* Returns zero.
*
*/
static int __init video_setup(char *options)
{
int i, global = 0;
if (!options || !*options)
global = 1;
if (!global && !strncmp(options, "ofonly", 6)) {
ofonly = 1;
global = 1;
}
if (!global && !strstr(options, "fb:")) {
global_mode_option = options;
global = 1;
}
if (!global) {
for (i = 0; i < FB_MAX; i++) {
if (video_options[i] == NULL) {
video_options[i] = options;
break;
}
}
}
return 1;
}
__setup("video=", video_setup);
#endif
/*
* Visible symbols for modules
*/
EXPORT_SYMBOL(register_framebuffer);
EXPORT_SYMBOL(unregister_framebuffer);
EXPORT_SYMBOL(num_registered_fb);
EXPORT_SYMBOL(registered_fb);
EXPORT_SYMBOL(fb_prepare_logo);
EXPORT_SYMBOL(fb_show_logo);
EXPORT_SYMBOL(fb_set_var);
EXPORT_SYMBOL(fb_blank);
EXPORT_SYMBOL(fb_pan_display);
EXPORT_SYMBOL(fb_get_buffer_offset);
EXPORT_SYMBOL(fb_set_suspend);
EXPORT_SYMBOL(fb_register_client);
EXPORT_SYMBOL(fb_unregister_client);
EXPORT_SYMBOL(fb_get_options);
MODULE_LICENSE("GPL");