Re: Fw: 2.6.8.1-mm2: floppy magically disappaperd

From: Bjorn Helgaas
Date: Fri Aug 20 2004 - 12:26:27 EST


> My flopped disappeared with this kernel becaouse of ACPI. It was
> working earlier revisions. Modprobing floppy.ko fails:
>
> (in dmesg)
> #v+
> inserting floppy driver for 2.6.8.1-mm2
> floppy: controller ACPI FDC0 at I/O 0x3f7-0x3f7 irq 6 dma channel 2
> Floppy drive(s): fd0 is 1.44M
> floppy0: Floppy io-port 0x03f7 in use
> #v-

Thanks for the bug report. Can you try the attached patch please?
This replaces the floppy.c patch from the -mm set.

I'm interested in the dmesg output whether it works or not.

I suspect your ACPI is reporting two I/O regions for the floppy
controller: 0x3f2-0x3f5, and 0x3f7. Strictly speaking, that's
probably the correct thing. It happens that my box (an HP DL360)
only reports 0x3f2-0x3f5. The driver uses 0x3f7, and I don't see
how it could get along without it, so I think the HP BIOS is broken.

In any case, the patch below should deal better with two I/O regions
being reported.

===== drivers/block/floppy.c 1.103 vs edited =====
--- 1.103/drivers/block/floppy.c 2004-08-02 02:00:45 -06:00
+++ edited/drivers/block/floppy.c 2004-08-20 11:04:56 -06:00
@@ -181,6 +181,13 @@
#include <linux/device.h>
#include <linux/buffer_head.h> /* for invalidate_buffers() */

+#ifdef CONFIG_ACPI_BUS
+#include <linux/acpi.h>
+#include <acpi/acpi_bus.h>
+
+static int no_acpi;
+#endif
+
/*
* PS/2 floppies have much slower step rates than regular floppies.
* It's been recommended that take about 1/4 of the default speed
@@ -4150,6 +4157,9 @@
{"slow", NULL, &slow_floppy, 1, 0},
{"unexpected_interrupts", NULL, &print_unex, 1, 0},
{"no_unexpected_interrupts", NULL, &print_unex, 0, 0},
+#ifdef CONFIG_ACPI_BUS
+ {"no_acpi", NULL, &no_acpi, 1, 0},
+#endif
{"L40SX", NULL, &print_unex, 0, 0}

EXTRA_FLOPPY_PARAMS
@@ -4222,10 +4232,152 @@
return get_disk(disks[drive]);
}

+#ifdef CONFIG_ACPI_BUS
+static int acpi_floppies;
+
+struct floppy_resources {
+ unsigned int port1;
+ unsigned int nr_ports1;
+ unsigned int port2;
+ unsigned int nr_ports2;
+ unsigned int irq;
+ unsigned int dma_channel;
+};
+
+static acpi_status acpi_floppy_resource(struct acpi_resource *res, void *data)
+{
+ struct floppy_resources *floppy = (struct floppy_resources *) data;
+ struct acpi_resource_io *io;
+ struct acpi_resource_irq *irq;
+ struct acpi_resource_ext_irq *ext_irq;
+ struct acpi_resource_dma *dma;
+
+ if (res->id == ACPI_RSTYPE_IO) {
+ io = &res->data.io;
+ if (io->range_length) {
+ printk("%s: %d ioports at 0x%x\n", __FUNCTION__,
+ io->range_length, io->min_base_address);
+ if (floppy->port1) {
+ floppy->port2 = io->min_base_address;
+ floppy->nr_ports2 = io->range_length;
+ } else {
+ floppy->port1 = io->min_base_address;
+ floppy->nr_ports1 = io->range_length;
+ }
+ }
+ } else if (res->id == ACPI_RSTYPE_IRQ) {
+ irq = &res->data.irq;
+ if (irq->number_of_interrupts > 0)
+ floppy->irq = acpi_register_gsi(irq->interrupts[0],
+ irq->edge_level, irq->active_high_low);
+ } else if (res->id == ACPI_RSTYPE_EXT_IRQ) {
+ ext_irq = &res->data.extended_irq;
+ if (ext_irq->number_of_interrupts > 0)
+ floppy->irq = acpi_register_gsi(ext_irq->interrupts[0],
+ ext_irq->edge_level, ext_irq->active_high_low);
+ } else if (res->id == ACPI_RSTYPE_DMA) {
+ dma = &res->data.dma;
+ if (dma->number_of_channels > 0)
+ floppy->dma_channel = dma->channels[0];
+ }
+ return AE_OK;
+}
+
+static int acpi_floppy_add(struct acpi_device *device)
+{
+ struct floppy_resources floppy;
+ acpi_status status;
+
+ memset(&floppy, 0, sizeof(floppy));
+ status = acpi_walk_resources(device->handle, METHOD_NAME__CRS,
+ acpi_floppy_resource, &floppy);
+ if (ACPI_FAILURE(status))
+ return -ENODEV;
+
+ printk("%s: controller ACPI %s at I/O 0x%x-0x%x",
+ DEVICE_NAME, device->pnp.bus_id,
+ floppy.port1, floppy.port1 + floppy.nr_ports1 - 1);
+ if (floppy.nr_ports2)
+ printk(", 0x%x-0x%x",
+ floppy.port2, floppy.port2 + floppy.nr_ports2 - 1);
+ printk(" irq %d dma channel %d\n", floppy.irq, floppy.dma_channel);
+
+ /*
+ * The driver assumes I/O port regions like 0x3f0-0x3f5 and 0x3f7,
+ * but it ignores the first two ports (i.e., 0x3f0 and 0x3f1),
+ * which are for PS/2 systems. Since no PS/2 systems have
+ * ACPI, we should get regions like 0x3f2-0x3f5 and 0x3f7.
+ * Some boxes don't report 0x3f7 at all, so try to make sense
+ * of whatever we get.
+ */
+ if (floppy.nr_ports1 == 4) { /* 0x3f2-0x3f5 */
+ floppy.port1 -= 2;
+ floppy.nr_ports1 = 6;
+ } else if (floppy.nr_ports1 == 6) { /* 0x3f2-0x3f7 */
+ floppy.port1 -= 2;
+ floppy.nr_ports1 = 6;
+ floppy.nr_ports2 = 1;
+ } else if (floppy.nr_ports1 == 8) { /* 0x3f0-0x3f7 */
+ floppy.nr_ports1 = 6;
+ floppy.nr_ports2 = 1;
+ }
+
+ if (floppy.nr_ports2 == 0)
+ printk(KERN_WARNING "%s: no FD_DIR/DCR port in %s _CRS, assuming 0x%x\n",
+ DEVICE_NAME, device->pnp.bus_id, floppy.port1 + 7);
+
+ if (acpi_floppies == 0) {
+ FDC1 = floppy.port1;
+ FLOPPY_IRQ = floppy.irq;
+ FLOPPY_DMA = floppy.dma_channel;
+ } else if (acpi_floppies == 1) {
+ FDC2 = floppy.port1;
+ if (floppy.irq != FLOPPY_IRQ || floppy.dma_channel != FLOPPY_DMA)
+ printk(KERN_WARNING "%s: different IRQ/DMA info for %s; may not work\n",
+ DEVICE_NAME, device->pnp.bus_id);
+ } else {
+ printk(KERN_ERR "%s: only 2 controllers supported; %s ignored\n",
+ DEVICE_NAME, device->pnp.bus_id);
+ return -ENODEV;
+ }
+
+ acpi_floppies++;
+ return 0;
+}
+
+static int acpi_floppy_remove(struct acpi_device *device, int type)
+{
+ printk(KERN_ERR "%s: remove ACPI %s not supported\n", DEVICE_NAME,
+ device->pnp.bus_id);
+ return -EINVAL;
+}
+
+static struct acpi_driver acpi_floppy_driver = {
+ .name = "floppy",
+ .ids = "PNP0700",
+ .ops = {
+ .add = acpi_floppy_add,
+ .remove = acpi_floppy_remove,
+ },
+};
+
+static int acpi_floppy_init(void)
+{
+ if (no_acpi)
+ return -ENODEV;
+ return acpi_bus_register_driver(&acpi_floppy_driver);
+}
+#endif
+
int __init floppy_init(void)
{
int i, unit, drive;
int err, dr;
+
+#ifdef CONFIG_ACPI_BUS
+ if (acpi_floppy_init() == 0)
+ return -ENODEV;
+#endif

raw_cmd = NULL;

-
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/