AcoSail/kernel_optimize_test
8
0
Fork 0
forked from luck/tmp_suning_uos_patched
Code Pull Requests Activity
03e22198d2
kernel_optimize_test/arch/x86/kernel/early-quirks.c
Thomas Gleixner 54ff7e595d x86: hpet: Work around hardware stupidity 
This more or less reverts commits 08be979 (x86: Force HPET
readback_cmp for all ATI chipsets) and 30a564be (x86, hpet: Restrict
read back to affected ATI chipsets) to the status of commit 8da854c
(x86, hpet: Erratum workaround for read after write of HPET
comparator).

The delta to commit 8da854c is mostly comments and the change from
WARN_ONCE to printk_once as we know the call path of this function
already.

This needs really in depth explanation:

First of all the HPET design is a complete failure. Having a counter
compare register which generates an interrupt on matching values
forces the software to do at least one superfluous readback of the
counter register.

While it is nice in theory to program "absolute" time events it is
practically useless because the timer runs at some absurd frequency
which can never be matched to real world units. So we are forced to
calculate a relative delta and this forces a readout of the actual
counter value, adding the delta and programming the compare
register. When the delta is small enough we run into the danger that
we program a compare value which is already in the past. Due to the
compare for equal nature of HPET we need to read back the counter
value after writing the compare rehgister (btw. this is necessary for
absolute timeouts as well) to make sure that we did not miss the timer
event. We try to work around that by setting the minimum delta to a
value which is larger than the theoretical time which elapses between
the counter readout and the compare register write, but that's only
true in theory. A NMI or SMI which hits between the readout and the
write can easily push us beyond that limit. This would result in
waiting for the next HPET timer interrupt until the 32bit wraparound
of the counter happens which takes about 306 seconds.

So we designed the next event function to look like:

   match = read_cnt() + delta;
   write_compare_ref(match);
   return read_cnt() < match ? 0 : -ETIME;

At some point we got into trouble with certain ATI chipsets. Even the
above "safe" procedure failed. The reason was that the write to the
compare register was delayed probably for performance reasons. The
theory was that they wanted to avoid the synchronization of the write
with the HPET clock, which is understandable. So the write does not
hit the compare register directly instead it goes to some intermediate
register which is copied to the real compare register in sync with the
HPET clock. That opens another window for hitting the dreaded "wait
for a wraparound" problem.

To work around that "optimization" we added a read back of the compare
register which either enforced the update of the just written value or
just delayed the readout of the counter enough to avoid the issue. We
unfortunately never got any affirmative info from ATI/AMD about this.

One thing is sure, that we nuked the performance "optimization" that
way completely and I'm pretty sure that the result is worse than
before some HW folks came up with those.

Just for paranoia reasons I added a check whether the read back
compare register value was the same as the value we wrote right
before. That paranoia check triggered a couple of years after it was
added on an Intel ICH9 chipset. Venki added a workaround (commit
8da854c) which was reading the compare register twice when the first
check failed. We considered this to be a penalty in general and
restricted the readback (thus the wasted CPU cycles) to the known to
be affected ATI chipsets.

This turned out to be a utterly wrong decision. 2.6.35 testers
experienced massive problems and finally one of them bisected it down
to commit 30a564be which spured some further investigation.

Finally we got confirmation that the write to the compare register can
be delayed by up to two HPET clock cycles which explains the problems
nicely. All we can do about this is to go back to Venki's initial
workaround in a slightly modified version.

Just for the record I need to say, that all of this could have been
avoided if hardware designers and of course the HPET committee would
have thought about the consequences for a split second. It's out of my
comprehension why designing a working timer is so hard. There are two
ways to achieve it:

 1) Use a counter wrap around aware compare_reg <= counter_reg
    implementation instead of the easy compare_reg == counter_reg

    Downsides:

	- It needs more silicon.

	- It needs a readout of the counter to apply a relative
	  timeout. This is necessary as the counter does not run in
	  any useful (and adjustable) frequency and there is no
	  guarantee that the counter which is used for timer events is
	  the same which is used for reading the actual time (and
	  therefor for calculating the delta)

    Upsides:

	- None

  2) Use a simple down counter for relative timer events

    Downsides:

	- Absolute timeouts are not possible, which is not a problem
	  at all in the context of an OS and the expected
	  max. latencies/jitter (also see Downsides of #1)

   Upsides:

	- It needs less or equal silicon.

	- It works ALWAYS

	- It is way faster than a compare register based solution (One
	  write versus one write plus at least one and up to four
	  reads)

I would not be so grumpy about all of this, if I would not have been
ignored for many years when pointing out these flaws to various
hardware folks. I really hate timers (at least those which seem to be
designed by janitors).

Though finally we got a reasonable explanation plus a solution and I
want to thank all the folks involved in chasing it down and providing
valuable input to this.

Bisected-by: Nix <nix@esperi.org.uk>
Reported-by: Artur Skawina <art.08.09@gmail.com>
Reported-by: Damien Wyart <damien.wyart@free.fr>
Reported-by: John Drescher <drescherjm@gmail.com>
Cc: Venkatesh Pallipadi <venki@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Andreas Herrmann <andreas.herrmann3@amd.com>
Cc: Borislav Petkov <borislav.petkov@amd.com>
Cc: stable@kernel.org
Acked-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2010-09-15 00:55:13 +02:00

292 lines
7.3 KiB
C
Raw Blame History

/* Various workarounds for chipset bugs.
This code runs very early and can't use the regular PCI subsystem
The entries are keyed to PCI bridges which usually identify chipsets
uniquely.
This is only for whole classes of chipsets with specific problems which
need early invasive action (e.g. before the timers are initialized).
Most PCI device specific workarounds can be done later and should be
in standard PCI quirks
Mainboard specific bugs should be handled by DMI entries.
CPU specific bugs in setup.c */
#include <linux/pci.h>
#include <linux/acpi.h>
#include <linux/pci_ids.h>
#include <asm/pci-direct.h>
#include <asm/dma.h>
#include <asm/io_apic.h>
#include <asm/apic.h>
#include <asm/iommu.h>
#include <asm/gart.h>
static void __init fix_hypertransport_config(int num, int slot, int func)
{
u32 htcfg;
/*
* we found a hypertransport bus
* make sure that we are broadcasting
* interrupts to all cpus on the ht bus
* if we're using extended apic ids
*/
htcfg = read_pci_config(num, slot, func, 0x68);
if (htcfg & (1 << 18)) {
printk(KERN_INFO "Detected use of extended apic ids "
"on hypertransport bus\n");
if ((htcfg & (1 << 17)) == 0) {
printk(KERN_INFO "Enabling hypertransport extended "
"apic interrupt broadcast\n");
printk(KERN_INFO "Note this is a bios bug, "
"please contact your hw vendor\n");
htcfg |= (1 << 17);
write_pci_config(num, slot, func, 0x68, htcfg);
}
}
}
static void __init via_bugs(int num, int slot, int func)
{
#ifdef CONFIG_GART_IOMMU
if ((max_pfn > MAX_DMA32_PFN || force_iommu) &&
!gart_iommu_aperture_allowed) {
printk(KERN_INFO
"Looks like a VIA chipset. Disabling IOMMU."
" Override with iommu=allowed\n");
gart_iommu_aperture_disabled = 1;
}
#endif
}
#ifdef CONFIG_ACPI
#ifdef CONFIG_X86_IO_APIC
static int __init nvidia_hpet_check(struct acpi_table_header *header)
{
return 0;
}
#endif /* CONFIG_X86_IO_APIC */
#endif /* CONFIG_ACPI */
static void __init nvidia_bugs(int num, int slot, int func)
{
#ifdef CONFIG_ACPI
#ifdef CONFIG_X86_IO_APIC
/*
* All timer overrides on Nvidia are
* wrong unless HPET is enabled.
* Unfortunately that's not true on many Asus boards.
* We don't know yet how to detect this automatically, but
* at least allow a command line override.
*/
if (acpi_use_timer_override)
return;
if (acpi_table_parse(ACPI_SIG_HPET, nvidia_hpet_check)) {
acpi_skip_timer_override = 1;
printk(KERN_INFO "Nvidia board "
"detected. Ignoring ACPI "
"timer override.\n");
printk(KERN_INFO "If you got timer trouble "
"try acpi_use_timer_override\n");
}
#endif
#endif
/* RED-PEN skip them on mptables too? */
}
#if defined(CONFIG_ACPI) && defined(CONFIG_X86_IO_APIC)
#if defined(CONFIG_ACPI) && defined(CONFIG_X86_IO_APIC)
static u32 __init ati_ixp4x0_rev(int num, int slot, int func)
{
u32 d;
u8 b;
b = read_pci_config_byte(num, slot, func, 0xac);
b &= ~(1<<5);
write_pci_config_byte(num, slot, func, 0xac, b);
d = read_pci_config(num, slot, func, 0x70);
d |= 1<<8;
write_pci_config(num, slot, func, 0x70, d);
d = read_pci_config(num, slot, func, 0x8);
d &= 0xff;
return d;
}
#endif
static void __init ati_bugs(int num, int slot, int func)
{
u32 d;
u8 b;
if (acpi_use_timer_override)
return;
d = ati_ixp4x0_rev(num, slot, func);
if (d < 0x82)
acpi_skip_timer_override = 1;
else {
/* check for IRQ0 interrupt swap */
outb(0x72, 0xcd6); b = inb(0xcd7);
if (!(b & 0x2))
acpi_skip_timer_override = 1;
}
if (acpi_skip_timer_override) {
printk(KERN_INFO "SB4X0 revision 0x%x\n", d);
printk(KERN_INFO "Ignoring ACPI timer override.\n");
printk(KERN_INFO "If you got timer trouble "
"try acpi_use_timer_override\n");
}
}
static u32 __init ati_sbx00_rev(int num, int slot, int func)
{
u32 old, d;
d = read_pci_config(num, slot, func, 0x70);
old = d;
d &= ~(1<<8);
write_pci_config(num, slot, func, 0x70, d);
d = read_pci_config(num, slot, func, 0x8);
d &= 0xff;
write_pci_config(num, slot, func, 0x70, old);
return d;
}
static void __init ati_bugs_contd(int num, int slot, int func)
{
u32 d, rev;
if (acpi_use_timer_override)
return;
rev = ati_sbx00_rev(num, slot, func);
if (rev > 0x13)
return;
/* check for IRQ0 interrupt swap */
d = read_pci_config(num, slot, func, 0x64);
if (!(d & (1<<14)))
acpi_skip_timer_override = 1;
if (acpi_skip_timer_override) {
printk(KERN_INFO "SB600 revision 0x%x\n", rev);
printk(KERN_INFO "Ignoring ACPI timer override.\n");
printk(KERN_INFO "If you got timer trouble "
"try acpi_use_timer_override\n");
}
}
#else
static void __init ati_bugs(int num, int slot, int func)
{
}
static void __init ati_bugs_contd(int num, int slot, int func)
{
}
#endif
#define QFLAG_APPLY_ONCE 0x1
#define QFLAG_APPLIED 0x2
#define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED)
struct chipset {
u32 vendor;
u32 device;
u32 class;
u32 class_mask;
u32 flags;
void (*f)(int num, int slot, int func);
};
/*
* Only works for devices on the root bus. If you add any devices
* not on bus 0 readd another loop level in early_quirks(). But
* be careful because at least the Nvidia quirk here relies on
* only matching on bus 0.
*/
static struct chipset early_qrk[] __initdata = {
{ PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID,
PCI_CLASS_BRIDGE_PCI, PCI_ANY_ID, QFLAG_APPLY_ONCE, nvidia_bugs },
{ PCI_VENDOR_ID_VIA, PCI_ANY_ID,
PCI_CLASS_BRIDGE_PCI, PCI_ANY_ID, QFLAG_APPLY_ONCE, via_bugs },
{ PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB,
PCI_CLASS_BRIDGE_HOST, PCI_ANY_ID, 0, fix_hypertransport_config },
{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP400_SMBUS,
PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs },
{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS,
PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd },
{}
};
/**
* check_dev_quirk - apply early quirks to a given PCI device
* @num: bus number
* @slot: slot number
* @func: PCI function
*
* Check the vendor & device ID against the early quirks table.
*
* If the device is single function, let early_quirks() know so we don't
* poke at this device again.
*/
static int __init check_dev_quirk(int num, int slot, int func)
{
u16 class;
u16 vendor;
u16 device;
u8 type;
int i;
class = read_pci_config_16(num, slot, func, PCI_CLASS_DEVICE);
if (class == 0xffff)
return -1; /* no class, treat as single function */
vendor = read_pci_config_16(num, slot, func, PCI_VENDOR_ID);
device = read_pci_config_16(num, slot, func, PCI_DEVICE_ID);
for (i = 0; early_qrk[i].f != NULL; i++) {
if (((early_qrk[i].vendor == PCI_ANY_ID) ||
(early_qrk[i].vendor == vendor)) &&
((early_qrk[i].device == PCI_ANY_ID) ||
(early_qrk[i].device == device)) &&
(!((early_qrk[i].class ^ class) &
early_qrk[i].class_mask))) {
if ((early_qrk[i].flags &
QFLAG_DONE) != QFLAG_DONE)
early_qrk[i].f(num, slot, func);
early_qrk[i].flags |= QFLAG_APPLIED;
}
}
type = read_pci_config_byte(num, slot, func,
PCI_HEADER_TYPE);
if (!(type & 0x80))
return -1;
return 0;
}
void __init early_quirks(void)
{
int slot, func;
if (!early_pci_allowed())
return;
/* Poor man's PCI discovery */
/* Only scan the root bus */
for (slot = 0; slot < 32; slot++)
for (func = 0; func < 8; func++) {
/* Only probe function 0 on single fn devices */
if (check_dev_quirk(0, slot, func))
break;
}
}
View Git Blame Copy Permalink