svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-09-03 Thread Tycho Nightingale
Author: tychon Date: Fri Apr 19 13:43:33 2019 New Revision: 346386 URL: https://svnweb.freebsd.org/changeset/base/346386 Log: remove the 4GB boundary requirement on PCI DMA segments Reviewed by: kib Discussed with: jhb Sponsored by: Dell EMC Isilon Differential Revision:

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread Tycho Nightingale
As Ryan suggests r232260 should be recommitted to get acc(4) fixed. However, given the age of the devices involved and the lack of support by the standard, I’d say the threshold isn’t met to reinstate the boundary globally preemptively. To get the insurance, which may not even be necessary, yo

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread Scott Long
Yeah, it might be turning into an old wives tale at this point. I clearly remember it being discussed at the PCI-SIG in late 2003 when PCIe 1.0 was in its final draft stages. However, that was a long time ago, and it’s possible that even if it’s a limitation in some version or another of the spe

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread Ryan Stone
+scottl@, who I believe explained this to us in the first place. As I recall, it had something to do with 64-bit DMA being expressed as segment base + 32-bit offset. DMA engines that blindly try to cross a 32-bit boundary end up back at the start of the segment and read/write the wrong memory loc

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread John Baldwin
I had looked for the aac change, but wasn't able to find it, perhaps because I looked at tags created in aac.c rather than aac_pci.c. I agree aac will need to be re-patched. I'm not really certain how many other devices are actually broken. They would all be due to a firmware bug, nothing inhere

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread Ryan Stone
Following up, this is what will have to be re-instated in the aac driver: http://svn.freebsd.org/changeset/base/232260 However, my biggest concern is that we have no idea how many new devices with the broken behaviour might have been introduced since we fixed the problem in general. How does Lin

Re: svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-25 Thread Ryan Stone
This change makes me *very* uncomfortable. It was originally brought in due to issues with Adaptec RAID cards using the aac(9) driver. The symptoms of the bug included silent corruption of data as it was written to disk. Are we sure that this change is a good idea, given how catastrophic it is w

svn commit: r346386 - in head/sys: dev/bge dev/pci dev/twa x86/iommu

2019-04-19 Thread Tycho Nightingale
Author: tychon Date: Fri Apr 19 13:43:33 2019 New Revision: 346386 URL: https://svnweb.freebsd.org/changeset/base/346386 Log: remove the 4GB boundary requirement on PCI DMA segments Reviewed by: kib Discussed with: jhb Sponsored by: Dell EMC Isilon Differential Revision: