linux-stable.git/arch/alpha/math-emu/math.c, branch linux-rolling-lts

alpha: don't leak hardware-fabricated FP exception bits to user space

2026-09-07T15:22:20+00:00

commit bcfe3187412e342b4619efb92c945f073855ebc0 upstream.

On EV6 and later the hardware records exception status bits in the FPCR
before delivering a software completion trap, and those bits can be wrong
for the instruction that trapped.  Converting a double that is exactly
representable as a subnormal float sets FPCR_UNF even though the result
is exact, and an underflow trap additionally sets FPCR_INE even when the
emulated operation turns out to be exact.

alpha_fp_emul() only wrote the FPCR when soft-fp raised an exception, so
whenever it determined that the instruction was exact the fabricated bits
stayed in the FPCR and were reported to user space by fetestexcept().

Pass the exception summary register down from do_entArith() so the
handler can tell which exceptions the hardware attributed to the trapping
instruction, and always write the FPCR.  Clear the exceptions that the
trap reported but that soft-fp did not raise.  EXC_SUM reports only the
underflow or overflow when the hardware also set INE, so treat INE as a
candidate in that case, and treat a trap with no reported exception as a
denormal operand trap, for which the hardware can fabricate INE and UNF
as well.  Bits that software has already confirmed in ieee_state belong
to this or an earlier instruction and are never cleared.

The imprecise path passes no summary.  There the trap was taken somewhere
in the trap shadow, so EXC_SUM is not attribution for the instruction
being re-executed -- and only EV6, which traps precisely and so never
takes that path, has fabricated bits to clear.  For the same reason the
clearing is guarded by implver(), matching swcr_update_status().

On an UP1500 (EV68) this takes the glibc math testsuite from 831 failures
to 28, the remainder being unrelated to exception status.

This belongs with the preceding fix to ieee_swcr_to_fpcr(), and should
not be backported without it -- nor it without this.  That fix stops
FPCR_DNOD being set unconditionally, so denormal operand traps start
firing again.  Those traps very often find an exact result, which is
precisely the case where the old code left the FPCR unwritten and the
fabricated bits visible.  Applied alone it would make spurious exception
flags more common, not less.

One case cannot be resolved here: an inexact instruction without the
software completion suffix never traps, so its INE reaches the FPCR
without being recorded anywhere else.  Such a bit is indistinguishable
from an INE the hardware fabricated for a trapping instruction, and is
lost if an underflow or overflow trap with an exact result follows it.
The FPCR is the only record of those instructions and it carries no
attribution.

The bug predates the git history, so there is no commit to reference in a
Fixes tag.

Cc: stable@vger.kernel.org # 5.15+
Signed-off-by: Matt Turner 
Reviewed-by: Magnus Lindholm 
Tested-by: Magnus Lindholm 
Link: https://lore.kernel.org/r/20260803-alpha-fp-exceptions-v1-2-c99d75608e60@gmail.com
Signed-off-by: Magnus Lindholm 
Signed-off-by: Greg Kroah-Hartman

alpha: missing includes

2024-05-03T20:09:30+00:00

... and missing externs in proto.h

Signed-off-by: Al Viro 
Acked-by: Paul E. McKenney 
Acked-by: Matt Turner 
Signed-off-by: Arnd Bergmann

alpha: move __udiv_qrnnd library function to arch/alpha/lib/

2021-09-18T21:45:48+00:00

We already had the implementation for __udiv_qrnnd (unsigned divide for
multi-precision arithmetic) as part of the alpha math emulation code.

But you can disable the math emulation code - even if you shouldn't -
and then the MPI code that actually wants this functionality (and is
needed by various crypto functions) will fail to build.

So move the extended-precision divide code to be a regular library
function, just like all the regular division code is.  That way ie is
available regardless of math-emulation.

Signed-off-by: Linus Torvalds

alpha: fp_emul: avoid init/cleanup_module names

2021-07-26T05:33:04+00:00

This is one of the last modules using the old calling conventions
for module init/exit functions. Change it over to the style used
everywhere else.

Signed-off-by: Arnd Bergmann 
Signed-off-by: Matt Turner

alpha: __udiv_qrnnd should be exported

2021-07-26T03:55:40+00:00

When building an alpha kernel with mpi set as module, I hit this build
error:
ERROR: "__udiv_qrnnd" [lib/mpi/mpi.ko] undefined!
make[2]: *** [scripts/Makefile.modpost:92: __modpost] Error 1
make[1]: *** [Makefile:1266: modules] Error 2

This is due to __udiv_qrnnd not exported.
Signed-off-by: Corentin Labbe 
Signed-off-by: Matt Turner

treewide: Add SPDX license identifier for more missed files

2019-05-21T08:50:45+00:00

Add SPDX license identifiers to all files which:

 - Have no license information of any form

 - Have MODULE_LICENCE("GPL*") inside which was used in the initial
   scan/conversion to ignore the file

These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:

  GPL-2.0-only

Signed-off-by: Thomas Gleixner 
Signed-off-by: Greg Kroah-Hartman

alpha: math-emu: Fix modular build

2017-09-04T19:04:34+00:00

Commit 00fc0e0dda62 ("alpha: move exports to actual definitions") also
removed the exports of the math emulator hooks, which are defined in C
code.  In case anyone cares about the option of CONFIG_MATHEMU=m, add
exports next to those definitions.  Also add a MODULE_LICENSE.

Fixes: 00fc0e0dda62 ("alpha: move exports to actual definitions")
Signed-off-by: Ben Hutchings 
Signed-off-by: Matt Turner

sched/headers: Prepare for the removal of from

2017-03-02T07:42:31+00:00

Fix up missing #includes in other places that rely on sched.h doing that for them.

Acked-by: Linus Torvalds 
Cc: Mike Galbraith 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar

Replace with globally

2016-12-24T19:46:01+00:00

This was entirely automated, using the script by Al:

  PATT='^[[:blank:]]*#[[:blank:]]*include[[:blank:]]*'
  sed -i -e "s!$PATT!#include !" \
        $(git grep -l "$PATT"|grep -v ^include/linux/uaccess.h)

to do the replacement at the end of the merge window.

Requested-by: Al Viro 
Signed-off-by: Linus Torvalds

alpha: fix conversion from denormal float to double

2008-01-17T23:38:59+00:00

The trap handler does properly update the fraction,
but not the exponent...

Thanks to Paolo Bonzini for the bug report and the testcase.

Signed-off-by: Ivan Kokshaysky 
Cc: Paolo Bonzini 
Cc: Richard Henderson 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds