Revision d7831a0bdf06b9f722b947bb0c205ff7d77cebd8 authored by Richard Kennedy on 30 June 2009, 18:41:35 UTC, committed by Linus Torvalds on 01 July 2009, 01:56:01 UTC
balance_dirty_pages can overreact and move all of the dirty pages to writeback unnecessarily. balance_dirty_pages makes its decision to throttle based on the number of dirty plus writeback pages that are over the calculated limit,so it will continue to move pages even when there are plenty of pages in writeback and less than the threshold still dirty. This allows it to overshoot its limits and move all the dirty pages to writeback while waiting for the drives to catch up and empty the writeback list. A simple fio test easily demonstrates this problem. fio --name=f1 --directory=/disk1 --size=2G -rw=write --name=f2 --directory=/disk2 --size=1G --rw=write --startdelay=10 This is the simplest fix I could find, but I'm not entirely sure that it alone will be enough for all cases. But it certainly is an improvement on my desktop machine writing to 2 disks. Do we need something more for machines with large arrays where bdi_threshold * number_of_drives is greater than the dirty_ratio ? Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Jens Axboe <jens.axboe@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent df279ca
o2micro.h
/*
* o2micro.h 1.13 1999/10/25 20:03:34
*
* The contents of this file are subject to the Mozilla Public License
* Version 1.1 (the "License"); you may not use this file except in
* compliance with the License. You may obtain a copy of the License
* at http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS"
* basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
* the License for the specific language governing rights and
* limitations under the License.
*
* The initial developer of the original code is David A. Hinds
* <dahinds@users.sourceforge.net>. Portions created by David A. Hinds
* are Copyright (C) 1999 David A. Hinds. All Rights Reserved.
*
* Alternatively, the contents of this file may be used under the
* terms of the GNU General Public License version 2 (the "GPL"), in which
* case the provisions of the GPL are applicable instead of the
* above. If you wish to allow the use of your version of this file
* only under the terms of the GPL and not to allow others to use
* your version of this file under the MPL, indicate your decision by
* deleting the provisions above and replace them with the notice and
* other provisions required by the GPL. If you do not delete the
* provisions above, a recipient may use your version of this file
* under either the MPL or the GPL.
*/
#ifndef _LINUX_O2MICRO_H
#define _LINUX_O2MICRO_H
#ifndef PCI_VENDOR_ID_O2
#define PCI_VENDOR_ID_O2 0x1217
#endif
#ifndef PCI_DEVICE_ID_O2_6729
#define PCI_DEVICE_ID_O2_6729 0x6729
#endif
#ifndef PCI_DEVICE_ID_O2_6730
#define PCI_DEVICE_ID_O2_6730 0x673a
#endif
#ifndef PCI_DEVICE_ID_O2_6832
#define PCI_DEVICE_ID_O2_6832 0x6832
#endif
#ifndef PCI_DEVICE_ID_O2_6836
#define PCI_DEVICE_ID_O2_6836 0x6836
#endif
#ifndef PCI_DEVICE_ID_O2_6812
#define PCI_DEVICE_ID_O2_6812 0x6872
#endif
/* Additional PCI configuration registers */
#define O2_MUX_CONTROL 0x90 /* 32 bit */
#define O2_MUX_RING_OUT 0x0000000f
#define O2_MUX_SKTB_ACTV 0x000000f0
#define O2_MUX_SCTA_ACTV_ENA 0x00000100
#define O2_MUX_SCTB_ACTV_ENA 0x00000200
#define O2_MUX_SER_IRQ_ROUTE 0x0000e000
#define O2_MUX_SER_PCI 0x00010000
#define O2_MUX_SKTA_TURBO 0x000c0000 /* for 6833, 6860 */
#define O2_MUX_SKTB_TURBO 0x00300000
#define O2_MUX_AUX_VCC_3V 0x00400000
#define O2_MUX_PCI_VCC_5V 0x00800000
#define O2_MUX_PME_MUX 0x0f000000
/* Additional ExCA registers */
#define O2_MODE_A 0x38
#define O2_MODE_A_2 0x26 /* for 6833B, 6860C */
#define O2_MODE_A_CD_PULSE 0x04
#define O2_MODE_A_SUSP_EDGE 0x08
#define O2_MODE_A_HOST_SUSP 0x10
#define O2_MODE_A_PWR_MASK 0x60
#define O2_MODE_A_QUIET 0x80
#define O2_MODE_B 0x39
#define O2_MODE_B_2 0x2e /* for 6833B, 6860C */
#define O2_MODE_B_IDENT 0x03
#define O2_MODE_B_ID_BSTEP 0x00
#define O2_MODE_B_ID_CSTEP 0x01
#define O2_MODE_B_ID_O2 0x02
#define O2_MODE_B_VS1 0x04
#define O2_MODE_B_VS2 0x08
#define O2_MODE_B_IRQ15_RI 0x80
#define O2_MODE_C 0x3a
#define O2_MODE_C_DREQ_MASK 0x03
#define O2_MODE_C_DREQ_INPACK 0x01
#define O2_MODE_C_DREQ_WP 0x02
#define O2_MODE_C_DREQ_BVD2 0x03
#define O2_MODE_C_ZVIDEO 0x08
#define O2_MODE_C_IREQ_SEL 0x30
#define O2_MODE_C_MGMT_SEL 0xc0
#define O2_MODE_D 0x3b
#define O2_MODE_D_IRQ_MODE 0x03
#define O2_MODE_D_PCI_CLKRUN 0x04
#define O2_MODE_D_CB_CLKRUN 0x08
#define O2_MODE_D_SKT_ACTV 0x20
#define O2_MODE_D_PCI_FIFO 0x40 /* for OZ6729, OZ6730 */
#define O2_MODE_D_W97_IRQ 0x40
#define O2_MODE_D_ISA_IRQ 0x80
#define O2_MHPG_DMA 0x3c
#define O2_MHPG_CHANNEL 0x07
#define O2_MHPG_CINT_ENA 0x08
#define O2_MHPG_CSC_ENA 0x10
#define O2_FIFO_ENA 0x3d
#define O2_FIFO_ZVIDEO_3 0x08
#define O2_FIFO_PCI_FIFO 0x10
#define O2_FIFO_POSTWR 0x40
#define O2_FIFO_BUFFER 0x80
#define O2_MODE_E 0x3e
#define O2_MODE_E_MHPG_DMA 0x01
#define O2_MODE_E_SPKR_OUT 0x02
#define O2_MODE_E_LED_OUT 0x08
#define O2_MODE_E_SKTA_ACTV 0x10
#define O2_RESERVED1 0x94
#define O2_RESERVED2 0xD4
#define O2_RES_READ_PREFETCH 0x02
#define O2_RES_WRITE_BURST 0x08
static int o2micro_override(struct yenta_socket *socket)
{
/*
* 'reserved' register at 0x94/D4. allows setting read prefetch and write
* bursting. read prefetching for example makes the RME Hammerfall DSP
* working. for some bridges it is at 0x94, for others at 0xD4. it's
* ok to write to both registers on all O2 bridges.
* from Eric Still, 02Micro.
*/
u8 a, b;
if (PCI_FUNC(socket->dev->devfn) == 0) {
a = config_readb(socket, O2_RESERVED1);
b = config_readb(socket, O2_RESERVED2);
dev_printk(KERN_INFO, &socket->dev->dev,
"O2: res at 0x94/0xD4: %02x/%02x\n", a, b);
switch (socket->dev->device) {
/*
* older bridges have problems with both read prefetch and write
* bursting depending on the combination of the chipset, bridge
* and the cardbus card. so disable them to be on the safe side.
*/
case PCI_DEVICE_ID_O2_6729:
case PCI_DEVICE_ID_O2_6730:
case PCI_DEVICE_ID_O2_6812:
case PCI_DEVICE_ID_O2_6832:
case PCI_DEVICE_ID_O2_6836:
dev_printk(KERN_INFO, &socket->dev->dev,
"Yenta O2: old bridge, disabling read "
"prefetch/write burst\n");
config_writeb(socket, O2_RESERVED1,
a & ~(O2_RES_READ_PREFETCH | O2_RES_WRITE_BURST));
config_writeb(socket, O2_RESERVED2,
b & ~(O2_RES_READ_PREFETCH | O2_RES_WRITE_BURST));
break;
default:
dev_printk(KERN_INFO , &socket->dev->dev,
"O2: enabling read prefetch/write burst\n");
config_writeb(socket, O2_RESERVED1,
a | O2_RES_READ_PREFETCH | O2_RES_WRITE_BURST);
config_writeb(socket, O2_RESERVED2,
b | O2_RES_READ_PREFETCH | O2_RES_WRITE_BURST);
}
}
return 0;
}
static void o2micro_restore_state(struct yenta_socket *socket)
{
/*
* as long as read prefetch is the only thing in
* o2micro_override, it's safe to call it from here
*/
o2micro_override(socket);
}
#endif /* _LINUX_O2MICRO_H */
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...