[Bug libc/27697] New: Technically-undefined shifts in sysdeps/x86/dl-cacheinfo.h

[Bug libc/27697] New: Technically-undefined shifts in sysdeps/x86/dl-cacheinfo.h

[brooks at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27697

            Bug ID: 27697
           Summary: Technically-undefined shifts in
                    sysdeps/x86/dl-cacheinfo.h
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: minor
          Priority: P2
         Component: libc
          Assignee: unassigned at sourceware dot org
          Reporter: brooks at gcc dot gnu.org
                CC: drepper.fsp at gmail dot com
  Target Milestone: ---

We recently got some positives on a "shift sanitizer" check, in
sysdeps/x86/dl-cacheinfo.h.

Specifically, at lines 611 and 628, we have:
    count_mask = ~(-1 << (count_mask + 1));

Technically, left shifts of negative numbers are undefined; we can avoid the
undefined behavior by respelling this as
    count_mask = ~(~0U << (count_mask + 1));

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/27697] Technically-undefined shifts in sysdeps/x86/dl-cacheinfo.h

[joseph at codesourcery dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27697

--- Comment #1 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
See <https://sourceware.org/legacy-ml/libc-alpha/2019-02/msg00177.html> - 
I failed to get consensus for that patch (my goal there was fixing 
warnings that appeared with -Wextra).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/27697] Technically-undefined shifts in sysdeps/x86/dl-cacheinfo.h

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27697

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fweimer at redhat dot com

--- Comment #2 from Florian Weimer <fweimer at redhat dot com> ---
GCC makes these shifts defined as an extension, but still flags them in the
sanitizers. So we are in a bit of an awkward situation here.

I think we should make the change if it helps with testing glibc with the
sanitizers. This piece of information wasn't available when the patch was
proposed the last time, I think.

-- 
You are receiving this mail because:
You are on the CC list for the bug.