android_kernel_xiaomi_sm8350/drivers/idle/i7300_idle.c
Tejun Heo 5a0e3ad6af include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files.  percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.

percpu.h -> slab.h dependency is about to be removed.  Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability.  As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.

  http://userweb.kernel.org/~tj/misc/slabh-sweep.py

The script does the followings.

* Scan files for gfp and slab usages and update includes such that
  only the necessary includes are there.  ie. if only gfp is used,
  gfp.h, if slab is used, slab.h.

* When the script inserts a new include, it looks at the include
  blocks and try to put the new include such that its order conforms
  to its surrounding.  It's put in the include block which contains
  core kernel includes, in the same order that the rest are ordered -
  alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
  doesn't seem to be any matching order.

* If the script can't find a place to put a new include (mostly
  because the file doesn't have fitting include block), it prints out
  an error message indicating which .h file needs to be added to the
  file.

The conversion was done in the following steps.

1. The initial automatic conversion of all .c files updated slightly
   over 4000 files, deleting around 700 includes and adding ~480 gfp.h
   and ~3000 slab.h inclusions.  The script emitted errors for ~400
   files.

2. Each error was manually checked.  Some didn't need the inclusion,
   some needed manual addition while adding it to implementation .h or
   embedding .c file was more appropriate for others.  This step added
   inclusions to around 150 files.

3. The script was run again and the output was compared to the edits
   from #2 to make sure no file was left behind.

4. Several build tests were done and a couple of problems were fixed.
   e.g. lib/decompress_*.c used malloc/free() wrappers around slab
   APIs requiring slab.h to be added manually.

5. The script was run on all .h files but without automatically
   editing them as sprinkling gfp.h and slab.h inclusions around .h
   files could easily lead to inclusion dependency hell.  Most gfp.h
   inclusion directives were ignored as stuff from gfp.h was usually
   wildly available and often used in preprocessor macros.  Each
   slab.h inclusion directive was examined and added manually as
   necessary.

6. percpu.h was updated not to include slab.h.

7. Build test were done on the following configurations and failures
   were fixed.  CONFIG_GCOV_KERNEL was turned off for all tests (as my
   distributed build env didn't work with gcov compiles) and a few
   more options had to be turned off depending on archs to make things
   build (like ipr on powerpc/64 which failed due to missing writeq).

   * x86 and x86_64 UP and SMP allmodconfig and a custom test config.
   * powerpc and powerpc64 SMP allmodconfig
   * sparc and sparc64 SMP allmodconfig
   * ia64 SMP allmodconfig
   * s390 SMP allmodconfig
   * alpha SMP allmodconfig
   * um on x86_64 SMP allmodconfig

8. percpu.h modifications were reverted so that it could be applied as
   a separate patch and serve as bisection point.

Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.

Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-30 22:02:32 +09:00

618 lines
16 KiB
C

/*
* (C) Copyright 2008 Intel Corporation
* Authors:
* Andy Henroid <andrew.d.henroid@intel.com>
* Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
*/
/*
* Save DIMM power on Intel 7300-based platforms when all CPUs/cores
* are idle, using the DIMM thermal throttling capability.
*
* This driver depends on the Intel integrated DMA controller (I/O AT).
* If the driver for I/O AT (drivers/dma/ioatdma*) is also enabled,
* this driver should work cooperatively.
*/
/* #define DEBUG */
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/gfp.h>
#include <linux/sched.h>
#include <linux/notifier.h>
#include <linux/cpumask.h>
#include <linux/ktime.h>
#include <linux/delay.h>
#include <linux/debugfs.h>
#include <linux/stop_machine.h>
#include <linux/i7300_idle.h>
#include <asm/idle.h>
#include "../dma/ioat/hw.h"
#include "../dma/ioat/registers.h"
#define I7300_IDLE_DRIVER_VERSION "1.55"
#define I7300_PRINT "i7300_idle:"
#define MAX_STOP_RETRIES 10
static int debug;
module_param_named(debug, debug, uint, 0644);
MODULE_PARM_DESC(debug, "Enable debug printks in this driver");
static int forceload;
module_param_named(forceload, forceload, uint, 0644);
MODULE_PARM_DESC(debug, "Enable driver testing on unvalidated i5000");
#define dprintk(fmt, arg...) \
do { if (debug) printk(KERN_INFO I7300_PRINT fmt, ##arg); } while (0)
/*
* Value to set THRTLOW to when initiating throttling
* 0 = No throttling
* 1 = Throttle when > 4 activations per eval window (Maximum throttling)
* 2 = Throttle when > 8 activations
* 168 = Throttle when > 672 activations (Minimum throttling)
*/
#define MAX_THROTTLE_LOW_LIMIT 168
static uint throttle_low_limit = 1;
module_param_named(throttle_low_limit, throttle_low_limit, uint, 0644);
MODULE_PARM_DESC(throttle_low_limit,
"Value for THRTLOWLM activation field "
"(0 = disable throttle, 1 = Max throttle, 168 = Min throttle)");
/*
* simple invocation and duration statistics
*/
static unsigned long total_starts;
static unsigned long total_us;
#ifdef DEBUG
static unsigned long past_skip;
#endif
static struct pci_dev *fbd_dev;
static spinlock_t i7300_idle_lock;
static int i7300_idle_active;
static u8 i7300_idle_thrtctl_saved;
static u8 i7300_idle_thrtlow_saved;
static u32 i7300_idle_mc_saved;
static cpumask_var_t idle_cpumask;
static ktime_t start_ktime;
static unsigned long avg_idle_us;
static struct dentry *debugfs_dir;
/* Begin: I/O AT Helper routines */
#define IOAT_CHANBASE(ioat_ctl, chan) (ioat_ctl + 0x80 + 0x80 * chan)
/* Snoop control (disable snoops when coherency is not important) */
#define IOAT_DESC_SADDR_SNP_CTL (1UL << 1)
#define IOAT_DESC_DADDR_SNP_CTL (1UL << 2)
static struct pci_dev *ioat_dev;
static struct ioat_dma_descriptor *ioat_desc; /* I/O AT desc & data (1 page) */
static unsigned long ioat_desc_phys;
static u8 *ioat_iomap; /* I/O AT memory-mapped control regs (aka CB_BAR) */
static u8 *ioat_chanbase;
/* Start I/O AT memory copy */
static int i7300_idle_ioat_start(void)
{
u32 err;
/* Clear error (due to circular descriptor pointer) */
err = readl(ioat_chanbase + IOAT_CHANERR_OFFSET);
if (err)
writel(err, ioat_chanbase + IOAT_CHANERR_OFFSET);
writeb(IOAT_CHANCMD_START, ioat_chanbase + IOAT1_CHANCMD_OFFSET);
return 0;
}
/* Stop I/O AT memory copy */
static void i7300_idle_ioat_stop(void)
{
int i;
u64 sts;
for (i = 0; i < MAX_STOP_RETRIES; i++) {
writeb(IOAT_CHANCMD_RESET,
ioat_chanbase + IOAT1_CHANCMD_OFFSET);
udelay(10);
sts = readq(ioat_chanbase + IOAT1_CHANSTS_OFFSET) &
IOAT_CHANSTS_STATUS;
if (sts != IOAT_CHANSTS_ACTIVE)
break;
}
if (i == MAX_STOP_RETRIES) {
dprintk("failed to stop I/O AT after %d retries\n",
MAX_STOP_RETRIES);
}
}
/* Test I/O AT by copying 1024 byte from 2k to 1k */
static int __init i7300_idle_ioat_selftest(u8 *ctl,
struct ioat_dma_descriptor *desc, unsigned long desc_phys)
{
u64 chan_sts;
memset(desc, 0, 2048);
memset((u8 *) desc + 2048, 0xab, 1024);
desc[0].size = 1024;
desc[0].ctl = 0;
desc[0].src_addr = desc_phys + 2048;
desc[0].dst_addr = desc_phys + 1024;
desc[0].next = 0;
writeb(IOAT_CHANCMD_RESET, ioat_chanbase + IOAT1_CHANCMD_OFFSET);
writeb(IOAT_CHANCMD_START, ioat_chanbase + IOAT1_CHANCMD_OFFSET);
udelay(1000);
chan_sts = readq(ioat_chanbase + IOAT1_CHANSTS_OFFSET) &
IOAT_CHANSTS_STATUS;
if (chan_sts != IOAT_CHANSTS_DONE) {
/* Not complete, reset the channel */
writeb(IOAT_CHANCMD_RESET,
ioat_chanbase + IOAT1_CHANCMD_OFFSET);
return -1;
}
if (*(u32 *) ((u8 *) desc + 3068) != 0xabababab ||
*(u32 *) ((u8 *) desc + 2044) != 0xabababab) {
dprintk("Data values src 0x%x, dest 0x%x, memset 0x%x\n",
*(u32 *) ((u8 *) desc + 2048),
*(u32 *) ((u8 *) desc + 1024),
*(u32 *) ((u8 *) desc + 3072));
return -1;
}
return 0;
}
static struct device dummy_dma_dev = {
.init_name = "fallback device",
.coherent_dma_mask = DMA_BIT_MASK(64),
.dma_mask = &dummy_dma_dev.coherent_dma_mask,
};
/* Setup and initialize I/O AT */
/* This driver needs I/O AT as the throttling takes effect only when there is
* some memory activity. We use I/O AT to set up a dummy copy, while all CPUs
* go idle and memory is throttled.
*/
static int __init i7300_idle_ioat_init(void)
{
u8 ver, chan_count, ioat_chan;
u16 chan_ctl;
ioat_iomap = (u8 *) ioremap_nocache(pci_resource_start(ioat_dev, 0),
pci_resource_len(ioat_dev, 0));
if (!ioat_iomap) {
printk(KERN_ERR I7300_PRINT "failed to map I/O AT registers\n");
goto err_ret;
}
ver = readb(ioat_iomap + IOAT_VER_OFFSET);
if (ver != IOAT_VER_1_2) {
printk(KERN_ERR I7300_PRINT "unknown I/O AT version (%u.%u)\n",
ver >> 4, ver & 0xf);
goto err_unmap;
}
chan_count = readb(ioat_iomap + IOAT_CHANCNT_OFFSET);
if (!chan_count) {
printk(KERN_ERR I7300_PRINT "unexpected # of I/O AT channels "
"(%u)\n",
chan_count);
goto err_unmap;
}
ioat_chan = chan_count - 1;
ioat_chanbase = IOAT_CHANBASE(ioat_iomap, ioat_chan);
chan_ctl = readw(ioat_chanbase + IOAT_CHANCTRL_OFFSET);
if (chan_ctl & IOAT_CHANCTRL_CHANNEL_IN_USE) {
printk(KERN_ERR I7300_PRINT "channel %d in use\n", ioat_chan);
goto err_unmap;
}
writew(IOAT_CHANCTRL_CHANNEL_IN_USE,
ioat_chanbase + IOAT_CHANCTRL_OFFSET);
ioat_desc = (struct ioat_dma_descriptor *)dma_alloc_coherent(
&dummy_dma_dev, 4096,
(dma_addr_t *)&ioat_desc_phys, GFP_KERNEL);
if (!ioat_desc) {
printk(KERN_ERR I7300_PRINT "failed to allocate I/O AT desc\n");
goto err_mark_unused;
}
writel(ioat_desc_phys & 0xffffffffUL,
ioat_chanbase + IOAT1_CHAINADDR_OFFSET_LOW);
writel(ioat_desc_phys >> 32,
ioat_chanbase + IOAT1_CHAINADDR_OFFSET_HIGH);
if (i7300_idle_ioat_selftest(ioat_iomap, ioat_desc, ioat_desc_phys)) {
printk(KERN_ERR I7300_PRINT "I/O AT self-test failed\n");
goto err_free;
}
/* Setup circular I/O AT descriptor chain */
ioat_desc[0].ctl = IOAT_DESC_SADDR_SNP_CTL | IOAT_DESC_DADDR_SNP_CTL;
ioat_desc[0].src_addr = ioat_desc_phys + 2048;
ioat_desc[0].dst_addr = ioat_desc_phys + 3072;
ioat_desc[0].size = 128;
ioat_desc[0].next = ioat_desc_phys + sizeof(struct ioat_dma_descriptor);
ioat_desc[1].ctl = ioat_desc[0].ctl;
ioat_desc[1].src_addr = ioat_desc[0].src_addr;
ioat_desc[1].dst_addr = ioat_desc[0].dst_addr;
ioat_desc[1].size = ioat_desc[0].size;
ioat_desc[1].next = ioat_desc_phys;
return 0;
err_free:
dma_free_coherent(&dummy_dma_dev, 4096, (void *)ioat_desc, 0);
err_mark_unused:
writew(0, ioat_chanbase + IOAT_CHANCTRL_OFFSET);
err_unmap:
iounmap(ioat_iomap);
err_ret:
return -ENODEV;
}
/* Cleanup I/O AT */
static void __exit i7300_idle_ioat_exit(void)
{
int i;
u64 chan_sts;
i7300_idle_ioat_stop();
/* Wait for a while for the channel to halt before releasing */
for (i = 0; i < MAX_STOP_RETRIES; i++) {
writeb(IOAT_CHANCMD_RESET,
ioat_chanbase + IOAT1_CHANCMD_OFFSET);
chan_sts = readq(ioat_chanbase + IOAT1_CHANSTS_OFFSET) &
IOAT_CHANSTS_STATUS;
if (chan_sts != IOAT_CHANSTS_ACTIVE) {
writew(0, ioat_chanbase + IOAT_CHANCTRL_OFFSET);
break;
}
udelay(1000);
}
chan_sts = readq(ioat_chanbase + IOAT1_CHANSTS_OFFSET) &
IOAT_CHANSTS_STATUS;
/*
* We tried to reset multiple times. If IO A/T channel is still active
* flag an error and return without cleanup. Memory leak is better
* than random corruption in that extreme error situation.
*/
if (chan_sts == IOAT_CHANSTS_ACTIVE) {
printk(KERN_ERR I7300_PRINT "Unable to stop IO A/T channels."
" Not freeing resources\n");
return;
}
dma_free_coherent(&dummy_dma_dev, 4096, (void *)ioat_desc, 0);
iounmap(ioat_iomap);
}
/* End: I/O AT Helper routines */
#define DIMM_THRTLOW 0x64
#define DIMM_THRTCTL 0x67
#define DIMM_THRTCTL_THRMHUNT (1UL << 0)
#define DIMM_MC 0x40
#define DIMM_GTW_MODE (1UL << 17)
#define DIMM_GBLACT 0x60
/*
* Keep track of an exponential-decaying average of recent idle durations.
* The latest duration gets DURATION_WEIGHT_PCT percentage weight
* in this average, with the old average getting the remaining weight.
*
* High weights emphasize recent history, low weights include long history.
*/
#define DURATION_WEIGHT_PCT 55
/*
* When the decaying average of recent durations or the predicted duration
* of the next timer interrupt is shorter than duration_threshold, the
* driver will decline to throttle.
*/
#define DURATION_THRESHOLD_US 100
/* Store DIMM thermal throttle configuration */
static int i7300_idle_thrt_save(void)
{
u32 new_mc_val;
u8 gblactlm;
pci_read_config_byte(fbd_dev, DIMM_THRTCTL, &i7300_idle_thrtctl_saved);
pci_read_config_byte(fbd_dev, DIMM_THRTLOW, &i7300_idle_thrtlow_saved);
pci_read_config_dword(fbd_dev, DIMM_MC, &i7300_idle_mc_saved);
/*
* Make sure we have Global Throttling Window Mode set to have a
* "short" window. This (mostly) works around an issue where
* throttling persists until the end of the global throttling window
* size. On the tested system, this was resulting in a maximum of
* 64 ms to exit throttling (average 32 ms). The actual numbers
* depends on system frequencies. Setting the short window reduces
* this by a factor of 4096.
*
* We will only do this only if the system is set for
* unlimited-activations while in open-loop throttling (i.e., when
* Global Activation Throttle Limit is zero).
*/
pci_read_config_byte(fbd_dev, DIMM_GBLACT, &gblactlm);
dprintk("thrtctl_saved = 0x%02x, thrtlow_saved = 0x%02x\n",
i7300_idle_thrtctl_saved,
i7300_idle_thrtlow_saved);
dprintk("mc_saved = 0x%08x, gblactlm = 0x%02x\n",
i7300_idle_mc_saved,
gblactlm);
if (gblactlm == 0) {
new_mc_val = i7300_idle_mc_saved | DIMM_GTW_MODE;
pci_write_config_dword(fbd_dev, DIMM_MC, new_mc_val);
return 0;
} else {
dprintk("could not set GTW_MODE = 1 (OLTT enabled)\n");
return -ENODEV;
}
}
/* Restore DIMM thermal throttle configuration */
static void i7300_idle_thrt_restore(void)
{
pci_write_config_dword(fbd_dev, DIMM_MC, i7300_idle_mc_saved);
pci_write_config_byte(fbd_dev, DIMM_THRTLOW, i7300_idle_thrtlow_saved);
pci_write_config_byte(fbd_dev, DIMM_THRTCTL, i7300_idle_thrtctl_saved);
}
/* Enable DIMM thermal throttling */
static void i7300_idle_start(void)
{
u8 new_ctl;
u8 limit;
new_ctl = i7300_idle_thrtctl_saved & ~DIMM_THRTCTL_THRMHUNT;
pci_write_config_byte(fbd_dev, DIMM_THRTCTL, new_ctl);
limit = throttle_low_limit;
if (unlikely(limit > MAX_THROTTLE_LOW_LIMIT))
limit = MAX_THROTTLE_LOW_LIMIT;
pci_write_config_byte(fbd_dev, DIMM_THRTLOW, limit);
new_ctl = i7300_idle_thrtctl_saved | DIMM_THRTCTL_THRMHUNT;
pci_write_config_byte(fbd_dev, DIMM_THRTCTL, new_ctl);
}
/* Disable DIMM thermal throttling */
static void i7300_idle_stop(void)
{
u8 new_ctl;
u8 got_ctl;
new_ctl = i7300_idle_thrtctl_saved & ~DIMM_THRTCTL_THRMHUNT;
pci_write_config_byte(fbd_dev, DIMM_THRTCTL, new_ctl);
pci_write_config_byte(fbd_dev, DIMM_THRTLOW, i7300_idle_thrtlow_saved);
pci_write_config_byte(fbd_dev, DIMM_THRTCTL, i7300_idle_thrtctl_saved);
pci_read_config_byte(fbd_dev, DIMM_THRTCTL, &got_ctl);
WARN_ON_ONCE(got_ctl != i7300_idle_thrtctl_saved);
}
/*
* i7300_avg_duration_check()
* return 0 if the decaying average of recent idle durations is
* more than DURATION_THRESHOLD_US
*/
static int i7300_avg_duration_check(void)
{
if (avg_idle_us >= DURATION_THRESHOLD_US)
return 0;
#ifdef DEBUG
past_skip++;
#endif
return 1;
}
/* Idle notifier to look at idle CPUs */
static int i7300_idle_notifier(struct notifier_block *nb, unsigned long val,
void *data)
{
unsigned long flags;
ktime_t now_ktime;
static ktime_t idle_begin_time;
static int time_init = 1;
if (!throttle_low_limit)
return 0;
if (unlikely(time_init)) {
time_init = 0;
idle_begin_time = ktime_get();
}
spin_lock_irqsave(&i7300_idle_lock, flags);
if (val == IDLE_START) {
cpumask_set_cpu(smp_processor_id(), idle_cpumask);
if (cpumask_weight(idle_cpumask) != num_online_cpus())
goto end;
now_ktime = ktime_get();
idle_begin_time = now_ktime;
if (i7300_avg_duration_check())
goto end;
i7300_idle_active = 1;
total_starts++;
start_ktime = now_ktime;
i7300_idle_start();
i7300_idle_ioat_start();
} else if (val == IDLE_END) {
cpumask_clear_cpu(smp_processor_id(), idle_cpumask);
if (cpumask_weight(idle_cpumask) == (num_online_cpus() - 1)) {
/* First CPU coming out of idle */
u64 idle_duration_us;
now_ktime = ktime_get();
idle_duration_us = ktime_to_us(ktime_sub
(now_ktime, idle_begin_time));
avg_idle_us =
((100 - DURATION_WEIGHT_PCT) * avg_idle_us +
DURATION_WEIGHT_PCT * idle_duration_us) / 100;
if (i7300_idle_active) {
ktime_t idle_ktime;
idle_ktime = ktime_sub(now_ktime, start_ktime);
total_us += ktime_to_us(idle_ktime);
i7300_idle_ioat_stop();
i7300_idle_stop();
i7300_idle_active = 0;
}
}
}
end:
spin_unlock_irqrestore(&i7300_idle_lock, flags);
return 0;
}
static struct notifier_block i7300_idle_nb = {
.notifier_call = i7300_idle_notifier,
};
MODULE_DEVICE_TABLE(pci, pci_tbl);
int stats_open_generic(struct inode *inode, struct file *fp)
{
fp->private_data = inode->i_private;
return 0;
}
static ssize_t stats_read_ul(struct file *fp, char __user *ubuf, size_t count,
loff_t *off)
{
unsigned long *p = fp->private_data;
char buf[32];
int len;
len = snprintf(buf, 32, "%lu\n", *p);
return simple_read_from_buffer(ubuf, count, off, buf, len);
}
static const struct file_operations idle_fops = {
.open = stats_open_generic,
.read = stats_read_ul,
};
struct debugfs_file_info {
void *ptr;
char name[32];
struct dentry *file;
} debugfs_file_list[] = {
{&total_starts, "total_starts", NULL},
{&total_us, "total_us", NULL},
#ifdef DEBUG
{&past_skip, "past_skip", NULL},
#endif
{NULL, "", NULL}
};
static int __init i7300_idle_init(void)
{
spin_lock_init(&i7300_idle_lock);
total_us = 0;
if (i7300_idle_platform_probe(&fbd_dev, &ioat_dev, forceload))
return -ENODEV;
if (i7300_idle_thrt_save())
return -ENODEV;
if (i7300_idle_ioat_init())
return -ENODEV;
if (!zalloc_cpumask_var(&idle_cpumask, GFP_KERNEL))
return -ENOMEM;
debugfs_dir = debugfs_create_dir("i7300_idle", NULL);
if (debugfs_dir) {
int i = 0;
while (debugfs_file_list[i].ptr != NULL) {
debugfs_file_list[i].file = debugfs_create_file(
debugfs_file_list[i].name,
S_IRUSR,
debugfs_dir,
debugfs_file_list[i].ptr,
&idle_fops);
i++;
}
}
idle_notifier_register(&i7300_idle_nb);
printk(KERN_INFO "i7300_idle: loaded v%s\n", I7300_IDLE_DRIVER_VERSION);
return 0;
}
static void __exit i7300_idle_exit(void)
{
idle_notifier_unregister(&i7300_idle_nb);
free_cpumask_var(idle_cpumask);
if (debugfs_dir) {
int i = 0;
while (debugfs_file_list[i].file != NULL) {
debugfs_remove(debugfs_file_list[i].file);
i++;
}
debugfs_remove(debugfs_dir);
}
i7300_idle_thrt_restore();
i7300_idle_ioat_exit();
}
module_init(i7300_idle_init);
module_exit(i7300_idle_exit);
MODULE_AUTHOR("Andy Henroid <andrew.d.henroid@intel.com>");
MODULE_DESCRIPTION("Intel Chipset DIMM Idle Power Saving Driver v"
I7300_IDLE_DRIVER_VERSION);
MODULE_LICENSE("GPL");