Revision 85bd839983778fcd0c1c043327b14a046e979b39 authored by Gu Zheng on 10 June 2015, 18:14:43 UTC, committed by Linus Torvalds on 10 June 2015, 23:43:43 UTC
Izumi found the following oops when hot re-adding a node: BUG: unable to handle kernel paging request at ffffc90008963690 IP: __wake_up_bit+0x20/0x70 Oops: 0000 [#1] SMP CPU: 68 PID: 1237 Comm: rs:main Q:Reg Not tainted 4.1.0-rc5 #80 Hardware name: FUJITSU PRIMEQUEST2800E/SB, BIOS PRIMEQUEST 2000 Series BIOS Version 1.87 04/28/2015 task: ffff880838df8000 ti: ffff880017b94000 task.ti: ffff880017b94000 RIP: 0010:[<ffffffff810dff80>] [<ffffffff810dff80>] __wake_up_bit+0x20/0x70 RSP: 0018:ffff880017b97be8 EFLAGS: 00010246 RAX: ffffc90008963690 RBX: 00000000003c0000 RCX: 000000000000a4c9 RDX: 0000000000000000 RSI: ffffea101bffd500 RDI: ffffc90008963648 RBP: ffff880017b97c08 R08: 0000000002000020 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000000 R12: ffff8a0797c73800 R13: ffffea101bffd500 R14: 0000000000000001 R15: 00000000003c0000 FS: 00007fcc7ffff700(0000) GS:ffff880874800000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffc90008963690 CR3: 0000000836761000 CR4: 00000000001407e0 Call Trace: unlock_page+0x6d/0x70 generic_write_end+0x53/0xb0 xfs_vm_write_end+0x29/0x80 [xfs] generic_perform_write+0x10a/0x1e0 xfs_file_buffered_aio_write+0x14d/0x3e0 [xfs] xfs_file_write_iter+0x79/0x120 [xfs] __vfs_write+0xd4/0x110 vfs_write+0xac/0x1c0 SyS_write+0x58/0xd0 system_call_fastpath+0x12/0x76 Code: 5d c3 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 55 48 89 e5 48 83 ec 20 65 48 8b 04 25 28 00 00 00 48 89 45 f8 31 c0 48 8d 47 48 <48> 39 47 48 48 c7 45 e8 00 00 00 00 48 c7 45 f0 00 00 00 00 48 RIP [<ffffffff810dff80>] __wake_up_bit+0x20/0x70 RSP <ffff880017b97be8> CR2: ffffc90008963690 Reproduce method (re-add a node):: Hot-add nodeA --> remove nodeA --> hot-add nodeA (panic) This seems an use-after-free problem, and the root cause is zone->wait_table was not set to *NULL* after free it in try_offline_node. When hot re-add a node, we will reuse the pgdat of it, so does the zone struct, and when add pages to the target zone, it will init the zone first (including the wait_table) if the zone is not initialized. The judgement of zone initialized is based on zone->wait_table: static inline bool zone_is_initialized(struct zone *zone) { return !!zone->wait_table; } so if we do not set the zone->wait_table to *NULL* after free it, the memory hotplug routine will skip the init of new zone when hot re-add the node, and the wait_table still points to the freed memory, then we will access the invalid address when trying to wake up the waiting people after the i/o operation with the page is done, such as mentioned above. Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com> Reported-by: Taku Izumi <izumi.taku@jp.fujitsu.com> Reviewed by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Tang Chen <tangchen@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent 5879ae5
tc86c001.c
/*
* Copyright (C) 2002 Toshiba Corporation
* Copyright (C) 2005-2006 MontaVista Software, Inc. <source@mvista.com>
*
* This file is licensed under the terms of the GNU General Public
* License version 2. This program is licensed "as is" without any
* warranty of any kind, whether express or implied.
*/
#include <linux/types.h>
#include <linux/pci.h>
#include <linux/ide.h>
#include <linux/module.h>
#define DRV_NAME "tc86c001"
static void tc86c001_set_mode(ide_hwif_t *hwif, ide_drive_t *drive)
{
unsigned long scr_port = hwif->config_data + (drive->dn ? 0x02 : 0x00);
u16 mode, scr = inw(scr_port);
const u8 speed = drive->dma_mode;
switch (speed) {
case XFER_UDMA_4: mode = 0x00c0; break;
case XFER_UDMA_3: mode = 0x00b0; break;
case XFER_UDMA_2: mode = 0x00a0; break;
case XFER_UDMA_1: mode = 0x0090; break;
case XFER_UDMA_0: mode = 0x0080; break;
case XFER_MW_DMA_2: mode = 0x0070; break;
case XFER_MW_DMA_1: mode = 0x0060; break;
case XFER_MW_DMA_0: mode = 0x0050; break;
case XFER_PIO_4: mode = 0x0400; break;
case XFER_PIO_3: mode = 0x0300; break;
case XFER_PIO_2: mode = 0x0200; break;
case XFER_PIO_1: mode = 0x0100; break;
case XFER_PIO_0:
default: mode = 0x0000; break;
}
scr &= (speed < XFER_MW_DMA_0) ? 0xf8ff : 0xff0f;
scr |= mode;
outw(scr, scr_port);
}
static void tc86c001_set_pio_mode(ide_hwif_t *hwif, ide_drive_t *drive)
{
drive->dma_mode = drive->pio_mode;
tc86c001_set_mode(hwif, drive);
}
/*
* HACKITY HACK
*
* This is a workaround for the limitation 5 of the TC86C001 IDE controller:
* if a DMA transfer terminates prematurely, the controller leaves the device's
* interrupt request (INTRQ) pending and does not generate a PCI interrupt (or
* set the interrupt bit in the DMA status register), thus no PCI interrupt
* will occur until a DMA transfer has been successfully completed.
*
* We work around this by initiating dummy, zero-length DMA transfer on
* a DMA timeout expiration. I found no better way to do this with the current
* IDE core than to temporarily replace a higher level driver's timer expiry
* handler with our own backing up to that handler in case our recovery fails.
*/
static int tc86c001_timer_expiry(ide_drive_t *drive)
{
ide_hwif_t *hwif = drive->hwif;
ide_expiry_t *expiry = ide_get_hwifdata(hwif);
u8 dma_stat = inb(hwif->dma_base + ATA_DMA_STATUS);
/* Restore a higher level driver's expiry handler first. */
hwif->expiry = expiry;
if ((dma_stat & 5) == 1) { /* DMA active and no interrupt */
unsigned long sc_base = hwif->config_data;
unsigned long twcr_port = sc_base + (drive->dn ? 0x06 : 0x04);
u8 dma_cmd = inb(hwif->dma_base + ATA_DMA_CMD);
printk(KERN_WARNING "%s: DMA interrupt possibly stuck, "
"attempting recovery...\n", drive->name);
/* Stop DMA */
outb(dma_cmd & ~0x01, hwif->dma_base + ATA_DMA_CMD);
/* Setup the dummy DMA transfer */
outw(0, sc_base + 0x0a); /* Sector Count */
outw(0, twcr_port); /* Transfer Word Count 1 or 2 */
/* Start the dummy DMA transfer */
/* clear R_OR_WCTR for write */
outb(0x00, hwif->dma_base + ATA_DMA_CMD);
/* set START_STOPBM */
outb(0x01, hwif->dma_base + ATA_DMA_CMD);
/*
* If an interrupt was pending, it should come thru shortly.
* If not, a higher level driver's expiry handler should
* eventually cause some kind of recovery from the DMA stall.
*/
return WAIT_MIN_SLEEP;
}
/* Chain to the restored expiry handler if DMA wasn't active. */
if (likely(expiry != NULL))
return expiry(drive);
/* If there was no handler, "emulate" that for ide_timer_expiry()... */
return -1;
}
static void tc86c001_dma_start(ide_drive_t *drive)
{
ide_hwif_t *hwif = drive->hwif;
unsigned long sc_base = hwif->config_data;
unsigned long twcr_port = sc_base + (drive->dn ? 0x06 : 0x04);
unsigned long nsectors = blk_rq_sectors(hwif->rq);
/*
* We have to manually load the sector count and size into
* the appropriate system control registers for DMA to work
* with LBA48 and ATAPI devices...
*/
outw(nsectors, sc_base + 0x0a); /* Sector Count */
outw(SECTOR_SIZE / 2, twcr_port); /* Transfer Word Count 1/2 */
/* Install our timeout expiry hook, saving the current handler... */
ide_set_hwifdata(hwif, hwif->expiry);
hwif->expiry = &tc86c001_timer_expiry;
ide_dma_start(drive);
}
static u8 tc86c001_cable_detect(ide_hwif_t *hwif)
{
struct pci_dev *dev = to_pci_dev(hwif->dev);
unsigned long sc_base = pci_resource_start(dev, 5);
u16 scr1 = inw(sc_base + 0x00);
/*
* System Control 1 Register bit 13 (PDIAGN):
* 0=80-pin cable, 1=40-pin cable
*/
return (scr1 & 0x2000) ? ATA_CBL_PATA40 : ATA_CBL_PATA80;
}
static void init_hwif_tc86c001(ide_hwif_t *hwif)
{
struct pci_dev *dev = to_pci_dev(hwif->dev);
unsigned long sc_base = pci_resource_start(dev, 5);
u16 scr1 = inw(sc_base + 0x00);
/* System Control 1 Register bit 15 (Soft Reset) set */
outw(scr1 | 0x8000, sc_base + 0x00);
/* System Control 1 Register bit 14 (FIFO Reset) set */
outw(scr1 | 0x4000, sc_base + 0x00);
/* System Control 1 Register: reset clear */
outw(scr1 & ~0xc000, sc_base + 0x00);
/* Store the system control register base for convenience... */
hwif->config_data = sc_base;
if (!hwif->dma_base)
return;
/*
* Sector Count Control Register bits 0 and 1 set:
* software sets Sector Count Register for master and slave device
*/
outw(0x0003, sc_base + 0x0c);
/* Sector Count Register limit */
hwif->rqsize = 0xffff;
}
static const struct ide_port_ops tc86c001_port_ops = {
.set_pio_mode = tc86c001_set_pio_mode,
.set_dma_mode = tc86c001_set_mode,
.cable_detect = tc86c001_cable_detect,
};
static const struct ide_dma_ops tc86c001_dma_ops = {
.dma_host_set = ide_dma_host_set,
.dma_setup = ide_dma_setup,
.dma_start = tc86c001_dma_start,
.dma_end = ide_dma_end,
.dma_test_irq = ide_dma_test_irq,
.dma_lost_irq = ide_dma_lost_irq,
.dma_timer_expiry = ide_dma_sff_timer_expiry,
.dma_sff_read_status = ide_dma_sff_read_status,
};
static const struct ide_port_info tc86c001_chipset = {
.name = DRV_NAME,
.init_hwif = init_hwif_tc86c001,
.port_ops = &tc86c001_port_ops,
.dma_ops = &tc86c001_dma_ops,
.host_flags = IDE_HFLAG_SINGLE | IDE_HFLAG_OFF_BOARD,
.pio_mask = ATA_PIO4,
.mwdma_mask = ATA_MWDMA2,
.udma_mask = ATA_UDMA4,
};
static int tc86c001_init_one(struct pci_dev *dev,
const struct pci_device_id *id)
{
int rc;
rc = pci_enable_device(dev);
if (rc)
goto out;
rc = pci_request_region(dev, 5, DRV_NAME);
if (rc) {
printk(KERN_ERR DRV_NAME ": system control regs already in use");
goto out_disable;
}
rc = ide_pci_init_one(dev, &tc86c001_chipset, NULL);
if (rc)
goto out_release;
goto out;
out_release:
pci_release_region(dev, 5);
out_disable:
pci_disable_device(dev);
out:
return rc;
}
static void tc86c001_remove(struct pci_dev *dev)
{
ide_pci_remove(dev);
pci_release_region(dev, 5);
pci_disable_device(dev);
}
static const struct pci_device_id tc86c001_pci_tbl[] = {
{ PCI_VDEVICE(TOSHIBA_2, PCI_DEVICE_ID_TOSHIBA_TC86C001_IDE), 0 },
{ 0, }
};
MODULE_DEVICE_TABLE(pci, tc86c001_pci_tbl);
static struct pci_driver tc86c001_pci_driver = {
.name = "TC86C001",
.id_table = tc86c001_pci_tbl,
.probe = tc86c001_init_one,
.remove = tc86c001_remove,
};
static int __init tc86c001_ide_init(void)
{
return ide_pci_register_driver(&tc86c001_pci_driver);
}
static void __exit tc86c001_ide_exit(void)
{
pci_unregister_driver(&tc86c001_pci_driver);
}
module_init(tc86c001_ide_init);
module_exit(tc86c001_ide_exit);
MODULE_AUTHOR("MontaVista Software, Inc. <source@mvista.com>");
MODULE_DESCRIPTION("PCI driver module for TC86C001 IDE");
MODULE_LICENSE("GPL");
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...