This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR33870, alias bug with SFTs and pointers
On Tue, 23 Oct 2007, Daniel Berlin wrote:
> On 10/23/07, Richard Guenther <rguenther@suse.de> wrote:
> >
> > In this bug we manage to prune SFTs for a subvariable because we
> > TBAA pruned the SFT for offset zero
> Why would we tbaa prune for offset 0. The TBAA sets for the struct
> should include the TBAA sets for all it's members.
The other testcase. The alias-15.c is for something I broke during
the fixing.
struct PgHdr {
unsigned int pgno;
PgHdr *pNextHash, *pPrevHash;
PgHdr *pNextFree, *pPrevFree;
PgHdr *pNextAll;
u8 inJournal;
short int nRef;
PgHdr *pDirty, *pPrevDirty;
unsigned int notUsed;
};
the NMT for which we enter add_virtual_operand through
get_indirect_ref_operands doesn't have SFT.6 in its aliases bitmap.
It's all in the audit trail of the PR if you want to look yourself.
Richard.