linux/drivers/ata, branch v2.6.35

ata_generic: implement ATA_GEN_* flags and force enable DMA on MBP 7,1

2010-07-01T19:34:48Z

IDE mode of MCP89 on MBP 7,1 doesn't set DMA enable bits in the BMDMA status register. Make the following changes to work around the problem. * Instead of using hard coded 1 in id->driver_data as class code match, use ATA_GEN_CLASS_MATCH and carry the matched id in host->private_data. * Instead of matching PCI_VENDOR_ID_CENATEK, use ATA_GEN_FORCE_DMA flag in id instead. * Add ATA_GEN_FORCE_DMA to the id entry of MBP 7,1. Signed-off-by: Tejun Heo Cc: Peer Chen Cc: stable@kernel.org Reported-by: Anders Østhus Reported-by: Andreas Graf Reported-by: Benoit Gschwind Reported-by: Damien Cassou Reported-by: tixetsal@juno.com Signed-off-by: Jeff Garzik

ahci,ata_generic: let ata_generic handle new MBP w/ MCP89

2010-07-01T19:34:46Z

For yet unknown reason, MCP89 on MBP 7,1 doesn't work w/ ahci under linux but the controller doesn't require explicit mode setting and works fine with ata_generic. Make ahci ignore the controller on MBP 7,1 and let ata_generic take it for now. Reported in bko#15923. https://bugzilla.kernel.org/show_bug.cgi?id=15923 NVIDIA is investigating why ahci mode doesn't work. Signed-off-by: Tejun Heo Cc: Peer Chen Cc: stable@kernel.org Reported-by: Anders Østhus Reported-by: Andreas Graf Reported-by: Benoit Gschwind Reported-by: Damien Cassou Reported-by: tixetsal@juno.com Signed-off-by: Jeff Garzik

libahci: Fix bug in storing EM messages

2010-07-01T19:34:15Z

In function ahci_store_em_buffer(), if the input (signed char*) buffer contains negative data, the constructed 32-bit long message data may be wrong. Signed-off-by: Harry Zhang Signed-off-by: Jeff Garzik

sata_sil24: Use memory barriers before issuing commands

2010-06-10T20:06:48Z

The data in the cmd_block buffers may reach the main memory after the writel() to the device ports. This patch introduces two calls to wmb() to ensure the relative ordering. Signed-off-by: Catalin Marinas Tested-by: Colin Tuckley Cc: Tejun Heo Cc: Jeff Garzik Signed-off-by: Jeff Garzik

sata_sil24: memset() overflow

2010-06-10T20:06:33Z

cb->atapi.cdb is an array of 16 u8 elements. The call too memset() would set the first part of the sge array to zero as well. It's not a packed struct. This one has been around for five years. I found it with Smatch. I think the reason no one has seen it before is because we normally call sil24_fill_sg() and that overwrites sge with proper information? Signed-off-by: Dan Carpenter Signed-off-by: Jeff Garzik

ahci: redo stopping DMA engines on empty ports

2010-06-07T20:03:15Z

Commit 96d60303fd (ahci: Turn off DMA engines when there's no device) implemented stopping DMA engines on empty ports but it used single sampling of status registers to determine device presence which led to disabling of DMA engines on occupied ports. Do it after all EH actions are complete using device presence state determined by EH. This avoids spurious disabling of DMA engines and simplifies the code. Signed-off-by: Tejun Heo Tested-by: Marc Dionne Cc: Matthew Garrett Cc: Robert Hancock Signed-off-by: Jeff Garzik

sata_sil24: fix kernel panic on ARM caused by unaligned access in sata_sil24

2010-06-07T20:03:13Z

The sata_sil24 driver has six 16-bit registers that are initialised with 32-bit writes. This cause a kernel panic on ARM due to the unaligned accesses which result. This patch changes the accesses to the correct 16-bit ones. Signed-off-by: Colin Tuckley Signed-off-by: Tejun Heo Signed-off-by: Jeff Garzik

sata_via: explain the magic fix

2010-06-07T20:03:08Z

Add Joseph Chan's explanation of the problem and workaround to the VT6421 magic fix. Signed-off-by: Tejun Heo Cc: Joseph Chan Signed-off-by: Jeff Garzik

Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev

2010-06-03T22:48:15Z

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev: libata: implement on-demand HPA unlocking libata: use the enlarged capacity after late HPA unlock SCSI: implement sd_unlock_native_capacity() libata-sff: trivial corrections to Kconfig help text sata_nv: don't diddle with nIEN on mcp55 sata_via: magic vt6421 fix for transmission problems w/ WD drives

libata: implement on-demand HPA unlocking

2010-06-02T17:50:10Z

Implement ata_scsi_unlock_native_capacity() which will be called through SCSI layer when block layer notices that partitions on a device extend beyond the end of the device. It requests EH to unlock HPA, waits for completion and returns the current device capacity. This allows libata to unlock HPA on demand instead of having to decide whether to unlock upfront. Unlocking on demand is safer than unlocking by upfront because some BIOSes write private data to the area beyond HPA limit. This was suggested by Ben Hutchings. Signed-off-by: Tejun Heo Suggested-by: Ben Hutchings Signed-off-by: Jeff Garzik