[Bug 236216] devel/valgrind: clang 8 crashes during build

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

[Bug 236216] devel/valgrind: clang 8 crashes during build


--- Comment #6 from [hidden email] ---
A commit references this bug:

Author: dim
Date: Thu Mar  7 19:33:40 UTC 2019
New revision: 344896
URL: https://svnweb.freebsd.org/changeset/base/344896

  ?Pull in r354937 from upstream clang trunk (by J?rg Sonnenberger):

    Fix inline assembler constraint validation

    The current constraint logic is both too lax and too strict. It fails
    for input outside the [INT_MIN..INT_MAX] range, but it also
    implicitly accepts 0 as value when it should not. Adjust logic to
    handle both correctly.

    Differential Revision: https://reviews.llvm.org/D58649

  Pull in r355491 from upstream clang trunk (by Hans Wennborg):

    Inline asm constraints: allow ICE-like pointers for the "n"
    constraint (PR40890)

    Apparently GCC allows this, and there's code relying on it (see bug).

    The idea is to allow expression that would have been allowed if they
    were cast to int. So I based the code on how such a cast would be
    done (the CK_PointerToIntegral case in

    Differential Revision: https://reviews.llvm.org/D58821

  These should fix assertions and errors when using the inline assembly
  "n" constraint in certain ways.

  In case of devel/valgrind, a pointer was used as the input for the
  constraint, which lead to "Assertion failed: (isInt() && "Invalid
  accessor"), function getInt".

  In case of math/secp256k1, a very large integer value was used as input
  for the constraint, which lead to "error: value '4624529908474429119'
  out of range for constraint 'n'".

  PR:             236216, 236194
  MFC after:      1 month
  X-MFC-With:     r344779


You are receiving this mail because:
You are the assignee for the bug.
[hidden email] mailing list
To unsubscribe, send any mail to "[hidden email]"