diff options
| author | Salyzyn, Mark <mark_salyzyn@adaptec.com> | 2006-06-26 08:37:39 -0400 |
|---|---|---|
| committer | James Bottomley <jejb@mulgrave.il.steeleye.com> | 2006-06-26 16:47:46 -0500 |
| commit | 12e9b5fb96028aab26af65bb06d2dee56c428a6e (patch) | |
| tree | 02208162fa852b78fc1908e8d97a1d866089b503 /drivers/scsi/aacraid | |
| parent | 99d19bb75b88cc997d3cd611903908714c735981 (diff) | |
| download | linux-3.10-12e9b5fb96028aab26af65bb06d2dee56c428a6e.tar.gz linux-3.10-12e9b5fb96028aab26af65bb06d2dee56c428a6e.tar.bz2 linux-3.10-12e9b5fb96028aab26af65bb06d2dee56c428a6e.zip | |
[SCSI] aacraid: remove x86_64 IOMMU dependent code
This may seem like a DILLIGAF, but after chatting with the F/W folks,
there is no harm in dropping the page calculation as denoted in the
enclosed patch for these older adapters in this new age of 4GB+ memory
sticks. Any resource optimization within the old-old-old adapters for
systems with less than 4G of memory is of little consequence. The
existing AAC_QUIRK_31BIT flag in linit.c should look after the rest of
the legacy hardware DMA limitations.
Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Diffstat (limited to 'drivers/scsi/aacraid')
| -rw-r--r-- | drivers/scsi/aacraid/comminit.c | 23 |
1 files changed, 1 insertions, 22 deletions
diff --git a/drivers/scsi/aacraid/comminit.c b/drivers/scsi/aacraid/comminit.c index 35b0a6ebd3f..1cd3584ba7f 100644 --- a/drivers/scsi/aacraid/comminit.c +++ b/drivers/scsi/aacraid/comminit.c @@ -92,28 +92,7 @@ static int aac_alloc_comm(struct aac_dev *dev, void **commaddr, unsigned long co init->AdapterFibsPhysicalAddress = cpu_to_le32((u32)phys); init->AdapterFibsSize = cpu_to_le32(fibsize); init->AdapterFibAlign = cpu_to_le32(sizeof(struct hw_fib)); - /* - * number of 4k pages of host physical memory. The aacraid fw needs - * this number to be less than 4gb worth of pages. num_physpages is in - * system page units. New firmware doesn't have any issues with the - * mapping system, but older Firmware did, and had *troubles* dealing - * with the math overloading past 32 bits, thus we must limit this - * field. - * - * This assumes the memory is mapped zero->n, which isnt - * always true on real computers. It also has some slight problems - * with the GART on x86-64. I've btw never tried DMA from PCI space - * on this platform but don't be surprised if its problematic. - */ -#ifndef CONFIG_GART_IOMMU - if ((num_physpages << (PAGE_SHIFT - 12)) <= AAC_MAX_HOSTPHYSMEMPAGES) { - init->HostPhysMemPages = - cpu_to_le32(num_physpages << (PAGE_SHIFT-12)); - } else -#endif - { - init->HostPhysMemPages = cpu_to_le32(AAC_MAX_HOSTPHYSMEMPAGES); - } + init->HostPhysMemPages = cpu_to_le32(AAC_MAX_HOSTPHYSMEMPAGES); init->InitFlags = 0; if (dev->new_comm_interface) { |