This is the mail archive of the
gcc-help@gcc.gnu.org
mailing list for the GCC project.
Re: GCC build in 64 bit OS failure
- From: Jeff Law <law at redhat dot com>
- To: Segher Boessenkool <segher at kernel dot crashing dot org>, vikram1729 <vsp1729 at gmail dot com>
- Cc: gcc-help at gcc dot gnu dot org
- Date: Mon, 4 Sep 2017 08:43:33 -0600
- Subject: Re: GCC build in 64 bit OS failure
- Authentication-results: sourceware.org; auth=none
- Authentication-results: ext-mx01.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com
- Authentication-results: ext-mx01.extmail.prod.ext.phx2.redhat.com; spf=fail smtp.mailfrom=law at redhat dot com
- Dmarc-filter: OpenDMARC Filter v1.3.2 mx1.redhat.com 7547E72F0C
- References: <1504524985158-0.post@n8.nabble.com> <20170904142110.GX13471@gate.crashing.org>
On 09/04/2017 08:21 AM, Segher Boessenkool wrote:
> Hi,
>
> On Mon, Sep 04, 2017 at 06:36:25AM -0500, vikram1729 wrote:
>> We are building GCC 4.1.2 cross toolchain for our custom processor in 64 bit
>> OS environment. Our custom processor is 32 bit. We faced with the following
>> ICE. This error did not come in 32 bit OS environment and toolchain build
>> successfully.
>
> That compiler is over ten years old. You might want to upgrade ;-)
>
>> t.c:3: error: unrecognizable insn:
>> (insn 22 21 23 (set:SI (reg:SI 19 r19)
>> (lo_sum:SI (reg:SI 19 r19)
>> (const_int 4294967280 [0xfffffff0]))) -1 (nil)
>> (nil))
>
> It looks like you meant -16 instead. Somewhere (could be anywhere
> really) you are treating the offset as an unsigned int instead of as
> a HOST_WIDE_INT?
>
> See where the wrong offset first appears (use -da for example), and
> work from there finding the problem.
>
> Good luck & have fun,
Yup. In RTL CONST_INTs are supposed to be sign extended.
jeff