It seems that on some nVidia controllers using AltStatus register can be unreliable so default to Status register if the PCI device is in Compatibility Mode. In order to achieve this: * Add ide_pci_is_in_compatibility_mode() inline helper to <linux/ide.h>. * Add IDE_HFLAG_BROKEN_ALTSTATUS host flag and set it in amd74xx host driver for nVidia controllers in Compatibility Mode. * Teach actual_try_to_identify() and drive_is_ready() about the new flag. This fixes the regression caused by removal of CONFIG_IDEPCI_SHARE_IRQ config option in 2.6.25 and using AltStatus register unconditionally when available (kernel.org bug #11659). [ Moreover for CONFIG_IDEPCI_SHARE_IRQ=y (which is what most distributions/people use) it never worked correctly. ] Thanks to Remy LABENE for help with debugging the problem. More info at: http://bugzilla.kernel.org/show_bug.cgi?id=11659 Reported-by: Remy LABENE <remy.labene@xxxxxxx> Cc: Borislav Petkov <petkovbb@xxxxxxxxxxxxxx> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@xxxxxxxxx> --- This fix should be safe for both .28 and -stable. drivers/ide/amd74xx.c | 11 ++++++++++- drivers/ide/ide-iops.c | 3 ++- drivers/ide/ide-probe.c | 3 ++- include/linux/ide.h | 8 ++++++++ 4 files changed, 22 insertions(+), 3 deletions(-) Index: b/drivers/ide/amd74xx.c =================================================================== --- a/drivers/ide/amd74xx.c +++ b/drivers/ide/amd74xx.c @@ -3,7 +3,7 @@ * IDE driver for Linux. * * Copyright (c) 2000-2002 Vojtech Pavlik - * Copyright (c) 2007 Bartlomiej Zolnierkiewicz + * Copyright (c) 2007-2008 Bartlomiej Zolnierkiewicz * * Based on the work of: * Andre Hedrick @@ -263,6 +263,15 @@ static int __devinit amd74xx_probe(struc d.udma_mask = ATA_UDMA5; } + /* + * It seems that on some nVidia controllers using AltStatus + * register can be unreliable so default to Status register + * if the device is in Compatibility Mode. + */ + if (dev->vendor == PCI_VENDOR_ID_NVIDIA && + ide_pci_is_in_compatibility_mode(dev)) + d.host_flags |= IDE_HFLAG_BROKEN_ALTSTATUS; + printk(KERN_INFO "%s %s: UDMA%s controller\n", d.name, pci_name(dev), amd_dma[fls(d.udma_mask) - 1]); Index: b/drivers/ide/ide-iops.c =================================================================== --- a/drivers/ide/ide-iops.c +++ b/drivers/ide/ide-iops.c @@ -468,7 +468,8 @@ int drive_is_ready (ide_drive_t *drive) * an interrupt with another pci card/device. We make no assumptions * about possible isa-pnp and pci-pnp issues yet. */ - if (hwif->io_ports.ctl_addr) + if (hwif->io_ports.ctl_addr && + (hwif->host_flags & IDE_HFLAG_BROKEN_ALTSTATUS) == 0) stat = hwif->tp_ops->read_altstatus(hwif); else /* Note: this may clear a pending IRQ!! */ Index: b/drivers/ide/ide-probe.c =================================================================== --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -266,7 +266,8 @@ static int actual_try_to_identify (ide_d /* take a deep breath */ msleep(50); - if (io_ports->ctl_addr) { + if (io_ports->ctl_addr && + (hwif->host_flags & IDE_HFLAG_BROKEN_ALTSTATUS) == 0) { a = tp_ops->read_altstatus(hwif); s = tp_ops->read_status(hwif); if ((a ^ s) & ~ATA_IDX) Index: b/include/linux/ide.h =================================================================== --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -1296,6 +1296,13 @@ extern int __ide_pci_register_driver(str #define ide_pci_register_driver(d) pci_register_driver(d) #endif +static inline int ide_pci_is_in_compatibility_mode(struct pci_dev *dev) +{ + if ((dev->class >> 8) == PCI_CLASS_STORAGE_IDE && (dev->class & 5) != 5) + return 1; + return 0; +} + void ide_pci_setup_ports(struct pci_dev *, const struct ide_port_info *, int, hw_regs_t *, hw_regs_t **); void ide_setup_pci_noise(struct pci_dev *, const struct ide_port_info *); @@ -1375,6 +1382,7 @@ enum { IDE_HFLAG_IO_32BIT = (1 << 24), /* unmask IRQs */ IDE_HFLAG_UNMASK_IRQS = (1 << 25), + IDE_HFLAG_BROKEN_ALTSTATUS = (1 << 26), /* serialize ports if DMA is possible (for sl82c105) */ IDE_HFLAG_SERIALIZE_DMA = (1 << 27), /* force host out of "simplex" mode */ -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
