[lkml]   [2007]   [Feb]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectUse resource_size_t for serial MMIO addresses
Andrew, please apply to -mm.  I think these should be good to merge
for 2.6.22.

At present, MMIO addresses for serial port (the ->mapbase field in
uart_port and other structures) are unsigned longs. This causes
problems on some 32-bit platforms which have a >32-bit physical
address bus, for example the embedded PowerPC 440GP chip, which has a
36-bit physical address bus, and on-chip serial ports located at an
MMIO address above 4GB.

The second patch in this series changes mapbase to a resource_size_t
(which can be 64-bit on the problematic platforms) in struct uart_port
and struct plat_serial8250_port. It does *not* change the type in
serial_struct, because that structure is exposed to userspace. It is
therefore unsafe to use setserial to change the address parameters on
a port using a mapbase above 4GB.

The first patch in the series contains the damage of the setserial
problem. It allows serial ports to be marked with a new
UPF_FIXED_PORT flag, which causes any attempts to change the port's
type (PIO/MMIO etc.), address or irq with setserial to be ignored.
While using setserial to alter the port address is useful for legacy
ISA ports, it is generally a bad idea for ports such as on-chip or
other hardwired ports where the arch code has good information (from
firmware or hardware probing) on the port's type and address.

David Gibson | I'll have my music baroque, and my code
david AT | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2007-02-20 04:19    [W:0.071 / U:1.196 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site