Re: Fix IDE locking error.

[Bartlomiej Zolnierkiewicz]

On 2/16/06, Dave Jones <davej@xxxxxxxxxx> wrote:
> This bit us a few kernels ago, and for some reason never made it's way
> upstream.

Because has never been submitted...

> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=144743
> Kernel panic - not syncing: drivers/ide/pci/piix.c:231:
> spin_lock(drivers/ide/ide.c:c03cef28) already locked by driver/ide/ide-iops.c/1153.

Could we get a decent description of the problem and of the patch?
Starting with the short description: it is piix locking being fixed not IDE one.

As this is piix only patch your bugzilla #144768 is not duplicate of
#144743.   Also some people reported the problem for atiixp.c
under #144743 (attixp.c has similar locking problem to piix.c).
So either they didn't care to reopen the bug or it was fixed by
some other change it the core IDE code.

> From: Alan Cox <alan@xxxxxxxxxx>
> Signed-off-by: Dave Jones <davej@xxxxxxxxxx>
>
> --- linux-2.6.12/drivers/ide/pci/piix.c~        2005-07-11 10:23:24.637181320 +0100
> +++ linux-2.6.12/drivers/ide/pci/piix.c 2005-07-11 10:23:24.637181320 +0100
> @@ -203,6 +203,8 @@
>         }
>  }
>
> +static spinlock_t tune_lock = SPIN_LOCK_UNLOCKED;
> +
>  /**
>   *     piix_tune_drive         -       tune a drive attached to a PIIX
>   *     @drive: drive to tune
> @@ -229,7 +231,12 @@
>                             { 2, 3 }, };
>
>         pio = ide_get_best_pio_mode(drive, pio, 5, NULL);
> -       spin_lock_irqsave(&ide_lock, flags);
> +
> +       /* Master v slave is synchronized above us but the slave register is
> +          shared by the two hwifs so the corner case of two slave timeouts in
> +          parallel must be locked */
> +
> +       spin_lock_irqsave(&tune_lock, flags);
>         pci_read_config_word(dev, master_port, &master_data);
>         if (is_slave) {
>                 master_data = master_data | 0x4000;
> @@ -249,7 +256,7 @@
>         pci_write_config_word(dev, master_port, master_data);
>         if (is_slave)
>                 pci_write_config_byte(dev, slave_port, slave_data);
> -       spin_unlock_irqrestore(&ide_lock, flags);
> +       spin_unlock_irqrestore(&tune_lock, flags);
>  }
>
>  /**
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/