The driver's prereset() method still doesn't check the channel enable bits. The bug was there for the entire time the driver has existed. :-/ Alan Cox fixed the HPT37x driver in commit b5bf24b94c65 ("[PATCH] hpt37x: Check the enablebits") but forgot to check the HPT3x2N driver which has the same bug. :-/ Signed-off-by: Sergey Shtylyov <s.shtylyov@xxxxxx> --- This patch is against the 'master' branch of Damien Le Moal's 'libata.git' repo. Changes in version 3: - added a passage to the patch description about Alan Cox missing to fix the driver back in 2006. Changes in version 2: - updated #define DRV_VERSION; - fixed typo in the patch description. drivers/ata/pata_hpt3x2n.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) Index: libata/drivers/ata/pata_hpt3x2n.c =================================================================== --- libata.orig/drivers/ata/pata_hpt3x2n.c +++ libata/drivers/ata/pata_hpt3x2n.c @@ -24,7 +24,7 @@ #include <linux/libata.h> #define DRV_NAME "pata_hpt3x2n" -#define DRV_VERSION "0.3.15" +#define DRV_VERSION "0.3.16" enum { HPT_PCI_FAST = (1 << 31), @@ -168,6 +168,13 @@ static int hpt3x2n_pre_reset(struct ata_ { struct ata_port *ap = link->ap; struct pci_dev *pdev = to_pci_dev(ap->host->dev); + static const struct pci_bits hpt3x2n_enable_bits[] = { + { 0x50, 1, 0x04, 0x04 }, + { 0x54, 1, 0x04, 0x04 } + }; + + if (!pci_test_config_bits(pdev, &hpt3x2n_enable_bits[ap->port_no])) + return -ENOENT; /* Reset the state machine */ pci_write_config_byte(pdev, 0x50 + 4 * ap->port_no, 0x37);