Re: RFC: Compact Flash True IDE Mode Driver

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

Hi,

On 2/1/06, Kumar Gala <galak@kernel.crashing.org> wrote:
> I was hoping to get some comments on this work in progress driver for
> using a compact flash device running in True IDE Mode connect via a MMIO
> interface.  The driver is working, however the embedded system I'm running
> on need some HW fixes to address the fact that the byte lanes for the data
> are swapped.  I figured now was a good time to incorporate any changes
> while I wait for the HW fixes (which will allow me to remove the
> cfide_outsw & cfide_insw).
>
> - kumar
>
>
> diff --git a/drivers/ide/Kconfig b/drivers/ide/Kconfig
> index 1c81174..8334baf 100644
> --- a/drivers/ide/Kconfig
> +++ b/drivers/ide/Kconfig
> @@ -930,6 +930,12 @@ config BLK_DEV_Q40IDE
>           normally be on; disable it only if you are running a custom hard
>           drive subsystem through an expansion card.
>
> +config BLK_DEV_CFIDE
> +       tristate "Compact Flash in True IDE mode"

depends on?

> +       help
> +         This is the IDE driver for a Memory Mapped Compact Flash Device
> +         running in True IDE mode.
> +
>  config BLK_DEV_MPC8xx_IDE
>         bool "MPC8xx IDE support"
>         depends on 8xx && IDE=y && BLK_DEV_IDE=y
> diff --git a/drivers/ide/Makefile b/drivers/ide/Makefile
> index 569fae7..09ed3aa 100644
> --- a/drivers/ide/Makefile
> +++ b/drivers/ide/Makefile
> @@ -32,6 +32,7 @@ ide-core-$(CONFIG_BLK_DEV_FALCON_IDE) +=
>  ide-core-$(CONFIG_BLK_DEV_GAYLE)       += legacy/gayle.o
>  ide-core-$(CONFIG_BLK_DEV_MAC_IDE)     += legacy/macide.o
>  ide-core-$(CONFIG_BLK_DEV_Q40IDE)      += legacy/q40ide.o
> +ide-core-$(CONFIG_BLK_DEV_CFIDE)       += legacy/cfide.o

Won't fly.  As comment in Makefile states "ide-core" is for
built-in only drivers.  Please see how ide-cs is handled.

>  # built-in only drivers from ppc/
>  ide-core-$(CONFIG_BLK_DEV_MPC8xx_IDE)  += ppc/mpc8xx.o
> diff --git a/drivers/ide/legacy/cfide.c b/drivers/ide/legacy/cfide.c
> new file mode 100644
> index 0000000..4444005
> --- /dev/null
> +++ b/drivers/ide/legacy/cfide.c
> @@ -0,0 +1,222 @@
> +/*
> + * Compact Flash Memory Mapped True IDE mode driver
> + *
> + * Maintainer: Kumar Gala <galak@kernel.crashing.org>
> + *
> + * This program is free software; you can redistribute  it and/or modify it
> + * under  the terms of  the GNU General  Public License as published by the
> + * Free Software Foundation;  either version 2 of the  License, or (at your
> + * option) any later version.
> + */
> +
> +#include <linux/config.h>

Is this needed?

> +#include <linux/types.h>
> +#include <linux/init.h>
> +#include <linux/kernel.h>
> +#include <linux/ide.h>
> +#include <linux/ioport.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <asm/io.h>
> +#include <asm/byteorder.h>

ditto

> +static void __iomem *cfide_mapbase;
> +static void __iomem *cfide_alt_mapbase;
> +
> +/* xxx: use standard outsw, insw when byte lanes swapped */
> +static void cfide_outsw(unsigned long port, void *addr, u32 count)
> +{
> +       _outsw((volatile u16 __iomem *)port, addr, count);
> +}
> +
> +static void cfide_insw(unsigned long port, void *addr, u32 count)
> +{
> +       _insw((volatile u16 __iomem *)port, addr, count);
> +}
> +
> +static void cfide_outl(u32 addr, unsigned long port)
> +{
> +       panic("outl unsupported");
> +}

Can be removed as it is not used in any code-path for this controller.

> +static void cfide_outsl(unsigned long port, void *addr, u32 count)
> +{
> +       panic("outsl unsupported");
> +}

This will panic as soon as somebody tries to enable 32-bit I/O
using hdparm.  Please add ide_hwif_t.no_io_32bit flag and teach
ide-disk.c:ide_disk_setup() about it (separate patch).

> +static u32 cfide_inl(unsigned long port)
> +{
> +       panic("inl unsupported");
> +       return 0;
> +}

same problem as with cfide_outl()

> +static void cfide_insl(unsigned long port, void *addr, u32 count)
> +{
> +       panic("insl unsupported");
> +}

same problem as with cfide_outsl()

> +static ide_hwif_t *cfide_locate_hwif(void __iomem * base, void __iomem * ctrl,
> +                                    unsigned int sz, int irq)

__devinit

> +{
> +       unsigned long port = (unsigned long)base;
> +       ide_hwif_t *hwif;
> +       int index, i;
> +
> +       for (index = 0; index < MAX_HWIFS; ++index) {
> +               hwif = ide_hwifs + index;
> +               if (hwif->io_ports[IDE_DATA_OFFSET] == port)
> +                       goto found;
> +       }
> +
> +       for (index = 0; index < MAX_HWIFS; ++index) {
> +               hwif = ide_hwifs + index;
> +               if (hwif->io_ports[IDE_DATA_OFFSET] == 0)
> +                       goto found;
> +       }
> +
> +       return NULL;

This is the same as in icside.c/rapide.c,
it really should be generic helper (separate patch)

> + found:
> +
> +       /* xxx: fix when byte lanes are swapped */
> +       hwif->hw.io_ports[IDE_DATA_OFFSET] = port;
> +       port += 3;
> +       for (i = IDE_ERROR_OFFSET; i <= IDE_STATUS_OFFSET; i++, port += sz)
> +               hwif->hw.io_ports[i] = port;
> +
> +       hwif->hw.io_ports[IDE_CONTROL_OFFSET] = (unsigned long)ctrl + 13;
> +
> +       memcpy(hwif->io_ports, hwif->hw.io_ports, sizeof(hwif->hw.io_ports));
> +       hwif->hw.irq = hwif->irq = irq;
> +
> +       hwif->hw.dma = NO_DMA;
> +       hwif->hw.chipset = ide_unknown;

Please use ide_generic or add new type,
ide_unknown indicates that this slot is free.

> +       hwif->mmio = 2;
> +       default_hwif_mmiops(hwif);
> +       hwif->OUTL = cfide_outl;

= NULL;

> +       hwif->OUTSW = cfide_outsw;
> +       hwif->OUTSL = cfide_outsl;

ditto

> +       hwif->INL = cfide_inl;

ditto

> +       hwif->INSW = cfide_insw;
> +       hwif->INSL = cfide_insl;

ditto

> +       return hwif;
> +}
> +
> +static int __devinit cfide_lbus_probe(struct platform_device *dev)
> +{
> +       struct resource *res_base, *res_alt, *res_irq;
> +       ide_hwif_t *hwif;
> +       int ret = 0;
> +
> +       /* get a pointer to the register memory */
> +       res_base = platform_get_resource(dev, IORESOURCE_MEM, 0);
> +       res_alt = platform_get_resource(dev, IORESOURCE_MEM, 1);
> +       res_irq = platform_get_resource(dev, IORESOURCE_IRQ, 0);
> +
> +       if ((!res_base) || (!res_alt) || (!res_irq)) {
> +               ret = -ENODEV;
> +               goto out;
> +       }
> +
> +       if (!request_mem_region
> +           (res_base->start, res_base->end - res_base->start + 1, dev->name)) {
> +               pr_debug("%s: request_mem_region of base failed\n", dev->name);
> +               ret = -EBUSY;
> +               goto out;
> +       }
> +
> +       if (!request_mem_region
> +           (res_alt->start, res_alt->end - res_alt->start + 1, dev->name)) {
> +               pr_debug("%s: request_mem_region of alt failed\n", dev->name);
> +               ret = -EBUSY;
> +               goto out;

res_base needs to be freed

> +       }
> +
> +       cfide_mapbase =
> +           ioremap(res_base->start, res_base->end - res_base->start + 1);
> +       if (!cfide_mapbase) {
> +               ret = -ENOMEM;
> +               goto out;

res_base and res_alt need to be freed

> +       }
> +
> +       cfide_alt_mapbase =
> +           ioremap(res_alt->start, res_alt->end - res_alt->start + 1);
> +
> +       if (!cfide_alt_mapbase) {
> +               ret = -ENOMEM;
> +               goto unmap_base;
> +       }
> +
> +       hwif = cfide_locate_hwif(cfide_mapbase, cfide_alt_mapbase, 2,
> +                             res_irq->start);
> +
> +       if (!hwif) {
> +               ret = -ENODEV;
> +               goto unmap_alt;
> +       }
> +
> +       hwif->gendev.parent = &dev->dev;
> +       hwif->noprobe = 0;
> +
> +       probe_hwif_init(hwif);

create_proc_ide_interfaces() is missing

> +       platform_set_drvdata(dev, hwif);
> +
> +       return 0;
> +
> + unmap_alt:
> +       iounmap(cfide_alt_mapbase);
> + unmap_base:
> +       iounmap(cfide_mapbase);
> + out:
> +       return ret;
> +}
> +
> +static int __devexit cfide_lbus_remove(struct platform_device *dev)
> +{
> +       ide_hwif_t *hwif = platform_get_drvdata(dev);
> +       struct resource *res_base, *res_alt;
> +
> +       /* get a pointer to the register memory */
> +       res_base = platform_get_resource(dev, IORESOURCE_MEM, 0);
> +       res_alt = platform_get_resource(dev, IORESOURCE_MEM, 1);
> +
> +       release_mem_region(res_base->start, res_base->end - res_base->start + 1);
> +       release_mem_region(res_alt->start, res_alt->end - res_alt->start + 1);
> +
> +       platform_set_drvdata(dev, NULL);
> +
> +       /* there must be a better way */

yes, fixing IDE unplug :)

> +       ide_unregister(hwif - ide_hwifs);

Otherwise driver looks pretty decent and with the above changes
it gets my ACK.  Additional enhancements as suggested by Ben
would be nice but obviously they are not needed for a start

Thanks,
Bartlomiej
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/