cvs commit: src/sys/dev/acpica acpi_pcib_acpi.c src/sys/i386/pci pci_bus.c

[M. Warner Losh]

In message: <432AFC3A.8000400 at root.org>
            Nate Lawson <nate at root.org> writes:
: M. Warner Losh wrote:
: > In message: <200509160702.j8G72TBv063544 at repoman.freebsd.org>
: >             Warner Losh <imp at FreeBSD.org> writes:
: > : imp         2005-09-16 07:02:29 UTC
: > : 
: > :   FreeBSD src repository
: > : 
: > :   Modified files:
: > :     sys/dev/acpica       acpi_pcib_acpi.c 
: > :     sys/i386/pci         pci_bus.c 
: > :   Log:
: > :   Commit a workaround to a problem with resource allocation.  This helps
: > :   with some Dell servers that booted w/o a problem[*] on 5.4, but failed
: > :   with 6.0-BETA.
: > :   
: > :   On the PCI bus, when we do lazy resource allocation, we narrow the
: > :   range requested as we pass through bridges to reflect how the bridges
: > :   are programmed and what addresses they pass.  However, when we're
: > :   doing an allocation on a bus that's directly connected to a host
: > :   bridge, no such translation can take place.  We already had a fallback
: > :   range for memory requests, but none for ioports.  As such, provide a
: > :   fallback for I/O ports so we don't allocate location 0, which will
: > :   have undesired side effects when the resources are actually used.
: > :   
: > :   This fixes a problem with booting a Dell server with usb in the
: > :   kernel.  However, it is an unsatisfying solution.  I don't like the
: > :   hard coded value, and I think we should start narrowing the resources
: > :   returned to not be in the so-called isa alias area (where the ranage &
: > :   0x0300 must be 0 iirc).  Doing such filtering will have to wait for
: > :   another day.
: > :   
: > :   This may be a good 6 candidate, maybe after its had a chance to be
: > :   refined.
: > :   
: > :   Tested by: glebius@
: > 
: > for those interested in the omitted footnote:
: > 
: > [*] and also without its first usb controller: uhci0 failed to attach.
: > 
: > I don't think this is worthy of a forced commit, but I know how some
: > people are when they see dangling references...
: 
: At some point, I may implement a similar workaround as Windows XP.  It 
: blocks access by AML to the following port ranges.
: 
: http://www.softwaretipsandtricks.com/forum/printthread.php?t=1674
: 
: Address       Function                 Comments
: 0x000 – 0x00F DMA Controller 1
: 0x020 – 0x021 PIC                      Access is never allowed*
: 0x040 – 0x043 System Timer 1
: 0x048 – 0x04B Timer 2 Failsafe
: 0x070 – 0x071 System CMOS, RTC
: 0x074 – 0x076 Extended CMOS
: 0x081 – 0x083 DMA1 Page Registers
: 0x087         DMA1 CH0 Low Page
: 0x089         DMA2 CH2 Low Page
: 0x08A – 0x08B DMA2 CH3 Low Page,
: 0x08F         DMA2 Low Page Refresh
: 0x090 – 0x091 Arbitration Control Port  Card Select Feedback
: 0x093 – 0x094 Reserved System Board Setup
: 0x096 – 0x097 POS Channel Select
: 0x0A0 – 0x0A1 Cascaded PIC              Access is never allowed*
: 0x0C0 – 0x0DF ISA DMA
: 0x4D0 – 0x4D1 PIC Edge/Level CR         Access is never allowed*
: 0xCF8 – 0xD00 PCI Configuration Space
: 
: *Read or write accesses to these ports are always blocked, regardless of 
: the BIOS use of the _OSI method.
: 
: Of course, you're describing device resources, but it seems like a 
: similar issue.  If someone else wants to do this, let me know.

I've been thinking for a long time of having silent devices that
gobble up these resources until such time that we have real devices.
Most of it could be driven by PnP or ACPI info, I believe.  The only
one that couldn't be is the PCI Config space, but one change to pci.c
would fix that :-).

Warner