0001 ===============================================
0002 Release notes for Linux Kernel VFP support code
0003 ===============================================
0004
0005 Date: 20 May 2004
0006
0007 Author: Russell King
0008
0009 This is the first release of the Linux Kernel VFP support code. It
0010 provides support for the exceptions bounced from VFP hardware found
0011 on ARM926EJ-S.
0012
0013 This release has been validated against the SoftFloat-2b library by
0014 John R. Hauser using the TestFloat-2a test suite. Details of this
0015 library and test suite can be found at:
0016
0017 http://www.jhauser.us/arithmetic/SoftFloat.html
0018
0019 The operations which have been tested with this package are:
0020
0021 - fdiv
0022 - fsub
0023 - fadd
0024 - fmul
0025 - fcmp
0026 - fcmpe
0027 - fcvtd
0028 - fcvts
0029 - fsito
0030 - ftosi
0031 - fsqrt
0032
0033 All the above pass softfloat tests with the following exceptions:
0034
0035 - fadd/fsub shows some differences in the handling of +0 / -0 results
0036 when input operands differ in signs.
0037 - the handling of underflow exceptions is slightly different. If a
0038 result underflows before rounding, but becomes a normalised number
0039 after rounding, we do not signal an underflow exception.
0040
0041 Other operations which have been tested by basic assembly-only tests
0042 are:
0043
0044 - fcpy
0045 - fabs
0046 - fneg
0047 - ftoui
0048 - ftosiz
0049 - ftouiz
0050
0051 The combination operations have not been tested:
0052
0053 - fmac
0054 - fnmac
0055 - fmsc
0056 - fnmsc
0057 - fnmul
