2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/proc_fs.h>
|
2006-01-11 13:17:48 -07:00
|
|
|
#include <linux/capability.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/ctype.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/seq_file.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
|
|
|
|
#define LINE_SIZE 80
|
|
|
|
|
|
|
|
#include <asm/mtrr.h>
|
|
|
|
#include "mtrr.h"
|
|
|
|
|
|
|
|
#define FILE_FCOUNT(f) (((struct seq_file *)((f)->private_data))->private)
|
|
|
|
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
static const char *const mtrr_strings[MTRR_NUM_TYPES] =
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
"uncachable", /* 0 */
|
|
|
|
"write-combining", /* 1 */
|
|
|
|
"?", /* 2 */
|
|
|
|
"?", /* 3 */
|
|
|
|
"write-through", /* 4 */
|
|
|
|
"write-protect", /* 5 */
|
|
|
|
"write-back", /* 6 */
|
|
|
|
};
|
|
|
|
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
const char *mtrr_attrib_to_str(int x)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
return (x <= 6) ? mtrr_strings[x] : "?";
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
|
|
|
static int
|
|
|
|
mtrr_file_add(unsigned long base, unsigned long size,
|
2008-01-30 05:30:31 -07:00
|
|
|
unsigned int type, bool increment, struct file *file, int page)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
int reg, max;
|
|
|
|
unsigned int *fcount = FILE_FCOUNT(file);
|
|
|
|
|
|
|
|
max = num_var_ranges;
|
|
|
|
if (fcount == NULL) {
|
2006-12-06 18:14:13 -07:00
|
|
|
fcount = kzalloc(max * sizeof *fcount, GFP_KERNEL);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!fcount)
|
|
|
|
return -ENOMEM;
|
|
|
|
FILE_FCOUNT(file) = fcount;
|
|
|
|
}
|
|
|
|
if (!page) {
|
|
|
|
if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1)))
|
|
|
|
return -EINVAL;
|
|
|
|
base >>= PAGE_SHIFT;
|
|
|
|
size >>= PAGE_SHIFT;
|
|
|
|
}
|
2008-01-30 05:30:31 -07:00
|
|
|
reg = mtrr_add_page(base, size, type, true);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (reg >= 0)
|
|
|
|
++fcount[reg];
|
|
|
|
return reg;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
mtrr_file_del(unsigned long base, unsigned long size,
|
|
|
|
struct file *file, int page)
|
|
|
|
{
|
|
|
|
int reg;
|
|
|
|
unsigned int *fcount = FILE_FCOUNT(file);
|
|
|
|
|
|
|
|
if (!page) {
|
|
|
|
if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1)))
|
|
|
|
return -EINVAL;
|
|
|
|
base >>= PAGE_SHIFT;
|
|
|
|
size >>= PAGE_SHIFT;
|
|
|
|
}
|
|
|
|
reg = mtrr_del_page(-1, base, size);
|
|
|
|
if (reg < 0)
|
|
|
|
return reg;
|
|
|
|
if (fcount == NULL)
|
|
|
|
return reg;
|
|
|
|
if (fcount[reg] < 1)
|
|
|
|
return -EINVAL;
|
|
|
|
--fcount[reg];
|
|
|
|
return reg;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* RED-PEN: seq_file can seek now. this is ignored. */
|
|
|
|
static ssize_t
|
|
|
|
mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos)
|
|
|
|
/* Format of control line:
|
|
|
|
"base=%Lx size=%Lx type=%s" OR:
|
|
|
|
"disable=%d"
|
|
|
|
*/
|
|
|
|
{
|
|
|
|
int i, err;
|
|
|
|
unsigned long reg;
|
|
|
|
unsigned long long base, size;
|
|
|
|
char *ptr;
|
|
|
|
char line[LINE_SIZE];
|
|
|
|
size_t linelen;
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
if (!len)
|
|
|
|
return -EINVAL;
|
|
|
|
memset(line, 0, LINE_SIZE);
|
|
|
|
if (len > LINE_SIZE)
|
|
|
|
len = LINE_SIZE;
|
|
|
|
if (copy_from_user(line, buf, len - 1))
|
|
|
|
return -EFAULT;
|
|
|
|
linelen = strlen(line);
|
|
|
|
ptr = line + linelen - 1;
|
|
|
|
if (linelen && *ptr == '\n')
|
|
|
|
*ptr = '\0';
|
|
|
|
if (!strncmp(line, "disable=", 8)) {
|
|
|
|
reg = simple_strtoul(line + 8, &ptr, 0);
|
|
|
|
err = mtrr_del_page(reg, 0, 0);
|
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
return len;
|
|
|
|
}
|
|
|
|
if (strncmp(line, "base=", 5))
|
|
|
|
return -EINVAL;
|
|
|
|
base = simple_strtoull(line + 5, &ptr, 0);
|
|
|
|
for (; isspace(*ptr); ++ptr) ;
|
|
|
|
if (strncmp(ptr, "size=", 5))
|
|
|
|
return -EINVAL;
|
|
|
|
size = simple_strtoull(ptr + 5, &ptr, 0);
|
|
|
|
if ((base & 0xfff) || (size & 0xfff))
|
|
|
|
return -EINVAL;
|
|
|
|
for (; isspace(*ptr); ++ptr) ;
|
|
|
|
if (strncmp(ptr, "type=", 5))
|
|
|
|
return -EINVAL;
|
|
|
|
ptr += 5;
|
|
|
|
for (; isspace(*ptr); ++ptr) ;
|
|
|
|
for (i = 0; i < MTRR_NUM_TYPES; ++i) {
|
|
|
|
if (strcmp(ptr, mtrr_strings[i]))
|
|
|
|
continue;
|
|
|
|
base >>= PAGE_SHIFT;
|
|
|
|
size >>= PAGE_SHIFT;
|
|
|
|
err =
|
|
|
|
mtrr_add_page((unsigned long) base, (unsigned long) size, i,
|
2008-01-30 05:30:31 -07:00
|
|
|
true);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
return len;
|
|
|
|
}
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2005-10-30 15:59:44 -07:00
|
|
|
static long
|
|
|
|
mtrr_ioctl(struct file *file, unsigned int cmd, unsigned long __arg)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2005-10-30 15:59:44 -07:00
|
|
|
int err = 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
mtrr_type type;
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
unsigned long size;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct mtrr_sentry sentry;
|
|
|
|
struct mtrr_gentry gentry;
|
|
|
|
void __user *arg = (void __user *) __arg;
|
|
|
|
|
2005-10-30 15:59:44 -07:00
|
|
|
switch (cmd) {
|
|
|
|
case MTRRIOC_ADD_ENTRY:
|
|
|
|
case MTRRIOC_SET_ENTRY:
|
|
|
|
case MTRRIOC_DEL_ENTRY:
|
|
|
|
case MTRRIOC_KILL_ENTRY:
|
|
|
|
case MTRRIOC_ADD_PAGE_ENTRY:
|
|
|
|
case MTRRIOC_SET_PAGE_ENTRY:
|
|
|
|
case MTRRIOC_DEL_PAGE_ENTRY:
|
|
|
|
case MTRRIOC_KILL_PAGE_ENTRY:
|
|
|
|
if (copy_from_user(&sentry, arg, sizeof sentry))
|
|
|
|
return -EFAULT;
|
|
|
|
break;
|
|
|
|
case MTRRIOC_GET_ENTRY:
|
|
|
|
case MTRRIOC_GET_PAGE_ENTRY:
|
|
|
|
if (copy_from_user(&gentry, arg, sizeof gentry))
|
|
|
|
return -EFAULT;
|
|
|
|
break;
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_ADD_ENTRY:
|
|
|
|
case MTRRIOC32_SET_ENTRY:
|
|
|
|
case MTRRIOC32_DEL_ENTRY:
|
|
|
|
case MTRRIOC32_KILL_ENTRY:
|
|
|
|
case MTRRIOC32_ADD_PAGE_ENTRY:
|
|
|
|
case MTRRIOC32_SET_PAGE_ENTRY:
|
|
|
|
case MTRRIOC32_DEL_PAGE_ENTRY:
|
|
|
|
case MTRRIOC32_KILL_PAGE_ENTRY: {
|
|
|
|
struct mtrr_sentry32 __user *s32 = (struct mtrr_sentry32 __user *)__arg;
|
|
|
|
err = get_user(sentry.base, &s32->base);
|
|
|
|
err |= get_user(sentry.size, &s32->size);
|
|
|
|
err |= get_user(sentry.type, &s32->type);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MTRRIOC32_GET_ENTRY:
|
|
|
|
case MTRRIOC32_GET_PAGE_ENTRY: {
|
|
|
|
struct mtrr_gentry32 __user *g32 = (struct mtrr_gentry32 __user *)__arg;
|
|
|
|
err = get_user(gentry.regnum, &g32->regnum);
|
|
|
|
err |= get_user(gentry.base, &g32->base);
|
|
|
|
err |= get_user(gentry.size, &g32->size);
|
|
|
|
err |= get_user(gentry.type, &g32->type);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
switch (cmd) {
|
|
|
|
default:
|
|
|
|
return -ENOTTY;
|
|
|
|
case MTRRIOC_ADD_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_ADD_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err =
|
2008-01-30 05:30:31 -07:00
|
|
|
mtrr_file_add(sentry.base, sentry.size, sentry.type, true,
|
2005-04-16 16:20:36 -06:00
|
|
|
file, 0);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_SET_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_SET_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
2008-01-30 05:30:31 -07:00
|
|
|
err = mtrr_add(sentry.base, sentry.size, sentry.type, false);
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
|
|
|
case MTRRIOC_DEL_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_DEL_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err = mtrr_file_del(sentry.base, sentry.size, file, 0);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_KILL_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_KILL_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err = mtrr_del(-1, sentry.base, sentry.size);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_GET_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_GET_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (gentry.regnum >= num_var_ranges)
|
|
|
|
return -EINVAL;
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
mtrr_if->get(gentry.regnum, &gentry.base, &size, &type);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/* Hide entries that go above 4GB */
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
if (gentry.base + size - 1 >= (1UL << (8 * sizeof(gentry.size) - PAGE_SHIFT))
|
|
|
|
|| size >= (1UL << (8 * sizeof(gentry.size) - PAGE_SHIFT)))
|
2005-04-16 16:20:36 -06:00
|
|
|
gentry.base = gentry.size = gentry.type = 0;
|
|
|
|
else {
|
|
|
|
gentry.base <<= PAGE_SHIFT;
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
gentry.size = size << PAGE_SHIFT;
|
2005-04-16 16:20:36 -06:00
|
|
|
gentry.type = type;
|
|
|
|
}
|
|
|
|
|
|
|
|
break;
|
|
|
|
case MTRRIOC_ADD_PAGE_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_ADD_PAGE_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err =
|
2008-01-30 05:30:31 -07:00
|
|
|
mtrr_file_add(sentry.base, sentry.size, sentry.type, true,
|
2005-04-16 16:20:36 -06:00
|
|
|
file, 1);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_SET_PAGE_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_SET_PAGE_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
2008-01-30 05:30:31 -07:00
|
|
|
err =
|
|
|
|
mtrr_add_page(sentry.base, sentry.size, sentry.type, false);
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
|
|
|
case MTRRIOC_DEL_PAGE_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_DEL_PAGE_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err = mtrr_file_del(sentry.base, sentry.size, file, 1);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_KILL_PAGE_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_KILL_PAGE_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
err = mtrr_del_page(-1, sentry.base, sentry.size);
|
|
|
|
break;
|
|
|
|
case MTRRIOC_GET_PAGE_ENTRY:
|
2007-02-13 05:26:26 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_GET_PAGE_ENTRY:
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
if (gentry.regnum >= num_var_ranges)
|
|
|
|
return -EINVAL;
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
mtrr_if->get(gentry.regnum, &gentry.base, &size, &type);
|
|
|
|
/* Hide entries that would overflow */
|
|
|
|
if (size != (__typeof__(gentry.size))size)
|
|
|
|
gentry.base = gentry.size = gentry.type = 0;
|
|
|
|
else {
|
|
|
|
gentry.size = size;
|
|
|
|
gentry.type = type;
|
|
|
|
}
|
2005-10-30 15:59:44 -07:00
|
|
|
break;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-10-30 15:59:44 -07:00
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
switch(cmd) {
|
|
|
|
case MTRRIOC_GET_ENTRY:
|
|
|
|
case MTRRIOC_GET_PAGE_ENTRY:
|
2005-04-16 16:20:36 -06:00
|
|
|
if (copy_to_user(arg, &gentry, sizeof gentry))
|
2005-10-30 15:59:44 -07:00
|
|
|
err = -EFAULT;
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
2005-10-30 15:59:44 -07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
case MTRRIOC32_GET_ENTRY:
|
|
|
|
case MTRRIOC32_GET_PAGE_ENTRY: {
|
|
|
|
struct mtrr_gentry32 __user *g32 = (struct mtrr_gentry32 __user *)__arg;
|
|
|
|
err = put_user(gentry.base, &g32->base);
|
|
|
|
err |= put_user(gentry.size, &g32->size);
|
|
|
|
err |= put_user(gentry.regnum, &g32->regnum);
|
|
|
|
err |= put_user(gentry.type, &g32->type);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2005-10-30 15:59:44 -07:00
|
|
|
return err;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
mtrr_close(struct inode *ino, struct file *file)
|
|
|
|
{
|
|
|
|
int i, max;
|
|
|
|
unsigned int *fcount = FILE_FCOUNT(file);
|
|
|
|
|
|
|
|
if (fcount != NULL) {
|
|
|
|
max = num_var_ranges;
|
|
|
|
for (i = 0; i < max; ++i) {
|
|
|
|
while (fcount[i] > 0) {
|
|
|
|
mtrr_del(i, 0, 0);
|
|
|
|
--fcount[i];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
kfree(fcount);
|
|
|
|
FILE_FCOUNT(file) = NULL;
|
|
|
|
}
|
|
|
|
return single_release(ino, file);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int mtrr_seq_show(struct seq_file *seq, void *offset);
|
|
|
|
|
|
|
|
static int mtrr_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
if (!mtrr_if)
|
|
|
|
return -EIO;
|
|
|
|
if (!mtrr_if->get)
|
|
|
|
return -ENXIO;
|
|
|
|
return single_open(file, mtrr_seq_show, NULL);
|
|
|
|
}
|
|
|
|
|
2007-02-12 01:55:31 -07:00
|
|
|
static const struct file_operations mtrr_fops = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = mtrr_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.write = mtrr_write,
|
2005-10-30 15:59:44 -07:00
|
|
|
.unlocked_ioctl = mtrr_ioctl,
|
|
|
|
.compat_ioctl = mtrr_ioctl,
|
2005-04-16 16:20:36 -06:00
|
|
|
.release = mtrr_close,
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
static struct proc_dir_entry *proc_root_mtrr;
|
|
|
|
|
|
|
|
|
|
|
|
static int mtrr_seq_show(struct seq_file *seq, void *offset)
|
|
|
|
{
|
|
|
|
char factor;
|
|
|
|
int i, max, len;
|
|
|
|
mtrr_type type;
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
unsigned long base, size;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
len = 0;
|
|
|
|
max = num_var_ranges;
|
|
|
|
for (i = 0; i < max; i++) {
|
|
|
|
mtrr_if->get(i, &base, &size, &type);
|
|
|
|
if (size == 0)
|
x86, 32-bit: trim memory not covered by wb mtrrs
On some machines, buggy BIOSes don't properly setup WB MTRRs to cover all
available RAM, meaning the last few megs (or even gigs) of memory will be
marked uncached. Since Linux tends to allocate from high memory addresses
first, this causes the machine to be unusably slow as soon as the kernel
starts really using memory (i.e. right around init time).
This patch works around the problem by scanning the MTRRs at boot and
figuring out whether the current end_pfn value (setup by early e820 code)
goes beyond the highest WB MTRR range, and if so, trimming it to match. A
fairly obnoxious KERN_WARNING is printed too, letting the user know that
not all of their memory is available due to a likely BIOS bug.
Something similar could be done on i386 if needed, but the boot ordering
would be slightly different, since the MTRR code on i386 depends on the
boot_cpu_data structure being setup.
This patch fixes a bug in the last patch that caused the code to run on
non-Intel machines (AMD machines apparently don't need it and it's untested
on other non-Intel machines, so best keep it off).
Further enhancements and fixes from:
Yinghai Lu <Yinghai.Lu@Sun.COM>
Andi Kleen <ak@suse.de>
Signed-off-by: Jesse Barnes <jesse.barnes@intel.com>
Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-01-30 05:33:18 -07:00
|
|
|
mtrr_usage_table[i] = 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
else {
|
|
|
|
if (size < (0x100000 >> PAGE_SHIFT)) {
|
|
|
|
/* less than 1MB */
|
|
|
|
factor = 'K';
|
|
|
|
size <<= PAGE_SHIFT - 10;
|
|
|
|
} else {
|
|
|
|
factor = 'M';
|
|
|
|
size >>= 20 - PAGE_SHIFT;
|
|
|
|
}
|
|
|
|
/* RED-PEN: base can be > 32bit */
|
|
|
|
len += seq_printf(seq,
|
[PATCH] i386: fix MTRR code
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2006-12-06 18:14:09 -07:00
|
|
|
"reg%02i: base=0x%05lx000 (%4luMB), size=%4lu%cB: %s, count=%d\n",
|
2005-04-16 16:20:36 -06:00
|
|
|
i, base, base >> (20 - PAGE_SHIFT), size, factor,
|
x86, 32-bit: trim memory not covered by wb mtrrs
On some machines, buggy BIOSes don't properly setup WB MTRRs to cover all
available RAM, meaning the last few megs (or even gigs) of memory will be
marked uncached. Since Linux tends to allocate from high memory addresses
first, this causes the machine to be unusably slow as soon as the kernel
starts really using memory (i.e. right around init time).
This patch works around the problem by scanning the MTRRs at boot and
figuring out whether the current end_pfn value (setup by early e820 code)
goes beyond the highest WB MTRR range, and if so, trimming it to match. A
fairly obnoxious KERN_WARNING is printed too, letting the user know that
not all of their memory is available due to a likely BIOS bug.
Something similar could be done on i386 if needed, but the boot ordering
would be slightly different, since the MTRR code on i386 depends on the
boot_cpu_data structure being setup.
This patch fixes a bug in the last patch that caused the code to run on
non-Intel machines (AMD machines apparently don't need it and it's untested
on other non-Intel machines, so best keep it off).
Further enhancements and fixes from:
Yinghai Lu <Yinghai.Lu@Sun.COM>
Andi Kleen <ak@suse.de>
Signed-off-by: Jesse Barnes <jesse.barnes@intel.com>
Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-01-30 05:33:18 -07:00
|
|
|
mtrr_attrib_to_str(type), mtrr_usage_table[i]);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __init mtrr_if_init(void)
|
|
|
|
{
|
|
|
|
struct cpuinfo_x86 *c = &boot_cpu_data;
|
|
|
|
|
|
|
|
if ((!cpu_has(c, X86_FEATURE_MTRR)) &&
|
|
|
|
(!cpu_has(c, X86_FEATURE_K6_MTRR)) &&
|
|
|
|
(!cpu_has(c, X86_FEATURE_CYRIX_ARR)) &&
|
|
|
|
(!cpu_has(c, X86_FEATURE_CENTAUR_MCR)))
|
|
|
|
return -ENODEV;
|
|
|
|
|
|
|
|
proc_root_mtrr =
|
|
|
|
create_proc_entry("mtrr", S_IWUSR | S_IRUGO, &proc_root);
|
|
|
|
if (proc_root_mtrr) {
|
|
|
|
proc_root_mtrr->owner = THIS_MODULE;
|
|
|
|
proc_root_mtrr->proc_fops = &mtrr_fops;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
arch_initcall(mtrr_if_init);
|
|
|
|
#endif /* CONFIG_PROC_FS */
|