https://bugs.kde.org/show_bug.cgi?id=514297

Mark Wielaard <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #36 from Mark Wielaard <[email protected]> ---
(In reply to mcermak from comment #34)
> Note that during rebasing my patch on top of master I've realized the
> following:
> - test might need to go to linux specific subdir

Yes, it should since only linux atm has guard pages.

> - there are couple of LINX_ --> LINXY replacements, because madvise now
> newly has POST(sys_madvise).  That however is linux specific, so some of
> these replacements happened also on non linux targets and those might need
> to be reverted

You mean GENX_(__NR_madvise, sys_madvise) going to GENX_(__NR_madvise,
sys_madvise)
also for syswrap-darwin and syswrap-freebsd. Yes, those are unnecessary and
could stay as GENX_.
But it is also mostly harmless since they are guarded with #ifdef
VKI_MADV_GUARD_INSTALL and
#ifdef VKI_MADV_GUARD_REMOVE. So the POST is just a noop on those systems.
Maybe slightly
inefficient, but probably unnoticeable. Lets ask Paul what he prefers (he has a
script that checks
that if there is a POST then there is also a Y).

> - finally, I've noticed this : https://paste.centos.org/view/d4b8b4c5 . it
> reads: *** File coregrind/m_aspacemgr/priv_aspacemgr.h must not include
> pub_tool_mallocfree.h   ... first time I see it

pub_tool is for tools, pub_core is for the coregrind code
(also you don't need a relative path for the include).
So that include code should be:

-#include "../pub_core_libcfile.h" // VG_(safe_fd)
-#include "pub_tool_mallocfree.h"  // VG_(calloc)
+#include "pub_core_libcfile.h"   // VG_(safe_fd)
+#include "pub_core_mallocfree.h" // VG_(calloc)

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to