Revision 0f640dca08330dfc7820d610578e5935b5e654b2 authored by Mike Snitzer on 31 January 2013, 14:11:14 UTC, committed by Alasdair G Kergon on 31 January 2013, 14:11:14 UTC
thin_io_hints() is blindly copying the queue limits from the thin-pool which can lead to incorrect limits being set. The fix here simply deletes the thin_io_hints() hook which leaves the existing stacking infrastructure to set the limits correctly. When a thin-pool uses an MD device for the data device a thin device from the thin-pool must respect MD's constraints about disallowing a bio from spanning multiple chunks. Otherwise we can see problems. If the raid0 chunksize is 1152K and thin-pool chunksize is 256K I see the following md/raid0 error (with extra debug tracing added to thin_endio) when mkfs.xfs is executed against the thin device: md/raid0:md99: make_request bug: can't convert block across chunks or bigger than 1152k 6688 127 device-mapper: thin: bio sector=2080 err=-5 bi_size=130560 bi_rw=17 bi_vcnt=32 bi_idx=0 This extra DM debugging shows that the failing bio is spanning across the first and second logical 1152K chunk (sector 2080 + 255 takes the bio beyond the first chunk's boundary of sector 2304). So the bio splitting that DM is doing clearly isn't respecting the MD limits. max_hw_sectors_kb is 127 for both the thin-pool and thin device (queue_max_hw_sectors returns 255 so we'll excuse sysfs's lack of precision). So this explains why bi_size is 130560. But the thin device's max_hw_sectors_kb should be 4 (PAGE_SIZE) given that it doesn't have a .merge function (for bio_add_page to consult indirectly via dm_merge_bvec) yet the thin-pool does sit above an MD device that has a compulsory merge_bvec_fn. This scenario is exactly why DM must resort to sending single PAGE_SIZE bios to the underlying layer. Some additional context for this is available in the header for commit 8cbeb67a ("dm: avoid unsupported spanning of md stripe boundaries"). Long story short, the reason a thin device doesn't properly get configured to have a max_hw_sectors_kb of 4 (PAGE_SIZE) is that thin_io_hints() is blindly copying the queue limits from the thin-pool device directly to the thin device's queue limits. Fix this by eliminating thin_io_hints. Doing so is safe because the block layer's queue limits stacking already enables the upper level thin device to inherit the thin-pool device's discard and minimum_io_size and optimal_io_size limits that get set in pool_io_hints. But avoiding the queue limits copy allows the thin and thin-pool limits to be different where it is important, namely max_hw_sectors_kb. Reported-by: Daniel Browning <db@kavod.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Cc: stable@vger.kernel.org Signed-off-by: Alasdair G Kergon <agk@redhat.com>
1 parent 949db15
autoprobe.c
/*
* linux/kernel/irq/autoprobe.c
*
* Copyright (C) 1992, 1998-2004 Linus Torvalds, Ingo Molnar
*
* This file contains the interrupt probing code and driver APIs.
*/
#include <linux/irq.h>
#include <linux/module.h>
#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/async.h>
#include "internals.h"
/*
* Autodetection depends on the fact that any interrupt that
* comes in on to an unassigned handler will get stuck with
* "IRQS_WAITING" cleared and the interrupt disabled.
*/
static DEFINE_MUTEX(probing_active);
/**
* probe_irq_on - begin an interrupt autodetect
*
* Commence probing for an interrupt. The interrupts are scanned
* and a mask of potential interrupt lines is returned.
*
*/
unsigned long probe_irq_on(void)
{
struct irq_desc *desc;
unsigned long mask = 0;
int i;
/*
* quiesce the kernel, or at least the asynchronous portion
*/
async_synchronize_full();
mutex_lock(&probing_active);
/*
* something may have generated an irq long ago and we want to
* flush such a longstanding irq before considering it as spurious.
*/
for_each_irq_desc_reverse(i, desc) {
raw_spin_lock_irq(&desc->lock);
if (!desc->action && irq_settings_can_probe(desc)) {
/*
* Some chips need to know about probing in
* progress:
*/
if (desc->irq_data.chip->irq_set_type)
desc->irq_data.chip->irq_set_type(&desc->irq_data,
IRQ_TYPE_PROBE);
irq_startup(desc, false);
}
raw_spin_unlock_irq(&desc->lock);
}
/* Wait for longstanding interrupts to trigger. */
msleep(20);
/*
* enable any unassigned irqs
* (we must startup again here because if a longstanding irq
* happened in the previous stage, it may have masked itself)
*/
for_each_irq_desc_reverse(i, desc) {
raw_spin_lock_irq(&desc->lock);
if (!desc->action && irq_settings_can_probe(desc)) {
desc->istate |= IRQS_AUTODETECT | IRQS_WAITING;
if (irq_startup(desc, false))
desc->istate |= IRQS_PENDING;
}
raw_spin_unlock_irq(&desc->lock);
}
/*
* Wait for spurious interrupts to trigger
*/
msleep(100);
/*
* Now filter out any obviously spurious interrupts
*/
for_each_irq_desc(i, desc) {
raw_spin_lock_irq(&desc->lock);
if (desc->istate & IRQS_AUTODETECT) {
/* It triggered already - consider it spurious. */
if (!(desc->istate & IRQS_WAITING)) {
desc->istate &= ~IRQS_AUTODETECT;
irq_shutdown(desc);
} else
if (i < 32)
mask |= 1 << i;
}
raw_spin_unlock_irq(&desc->lock);
}
return mask;
}
EXPORT_SYMBOL(probe_irq_on);
/**
* probe_irq_mask - scan a bitmap of interrupt lines
* @val: mask of interrupts to consider
*
* Scan the interrupt lines and return a bitmap of active
* autodetect interrupts. The interrupt probe logic state
* is then returned to its previous value.
*
* Note: we need to scan all the irq's even though we will
* only return autodetect irq numbers - just so that we reset
* them all to a known state.
*/
unsigned int probe_irq_mask(unsigned long val)
{
unsigned int mask = 0;
struct irq_desc *desc;
int i;
for_each_irq_desc(i, desc) {
raw_spin_lock_irq(&desc->lock);
if (desc->istate & IRQS_AUTODETECT) {
if (i < 16 && !(desc->istate & IRQS_WAITING))
mask |= 1 << i;
desc->istate &= ~IRQS_AUTODETECT;
irq_shutdown(desc);
}
raw_spin_unlock_irq(&desc->lock);
}
mutex_unlock(&probing_active);
return mask & val;
}
EXPORT_SYMBOL(probe_irq_mask);
/**
* probe_irq_off - end an interrupt autodetect
* @val: mask of potential interrupts (unused)
*
* Scans the unused interrupt lines and returns the line which
* appears to have triggered the interrupt. If no interrupt was
* found then zero is returned. If more than one interrupt is
* found then minus the first candidate is returned to indicate
* their is doubt.
*
* The interrupt probe logic state is returned to its previous
* value.
*
* BUGS: When used in a module (which arguably shouldn't happen)
* nothing prevents two IRQ probe callers from overlapping. The
* results of this are non-optimal.
*/
int probe_irq_off(unsigned long val)
{
int i, irq_found = 0, nr_of_irqs = 0;
struct irq_desc *desc;
for_each_irq_desc(i, desc) {
raw_spin_lock_irq(&desc->lock);
if (desc->istate & IRQS_AUTODETECT) {
if (!(desc->istate & IRQS_WAITING)) {
if (!nr_of_irqs)
irq_found = i;
nr_of_irqs++;
}
desc->istate &= ~IRQS_AUTODETECT;
irq_shutdown(desc);
}
raw_spin_unlock_irq(&desc->lock);
}
mutex_unlock(&probing_active);
if (nr_of_irqs > 1)
irq_found = -irq_found;
return irq_found;
}
EXPORT_SYMBOL(probe_irq_off);
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...