Re: Need help debugging silo
On 29/12/15 11:03, John Paul Adrian Glaubitz wrote:
> On 12/29/2015 11:52 AM, Richard Mortimer wrote:
>> From memory the conversion to a.out is done by the elftoaout binary from
>> sparc-utils. It maybe that the problem lies in there rather than in silo
>> itself.
> Right!
>
> That must be it:
>
> typedef struct {
> unsigned long a_magic; /* magic number */
> unsigned long a_text; /* size of text segment */
> unsigned long a_data; /* size of initialized data */
> unsigned long a_bss; /* size of uninitialized data */
> unsigned long a_syms; /* size of symbol table ||
> checksum */
> unsigned long a_entry; /* entry point */
> unsigned long a_trsize; /* size of text relocation */
> unsigned long a_drsize; /* size of data relocation */
> } Exec;
>
> The unsigned long probably becomes 64 bit on sparc64. I'll check whether
> Oracle has a patched version.
>
> Adrian
>
(yawn) morning or something similar,..
I used elftoaout-2.3-14.el6.sparc64 which I think was unmodified from the f12
archives.. -checking- yes it was direct from
https://kojipkgs.fedoraproject.org//packages/elftoaout/2.3/14.fc12/src/elftoaout-2.3-14.fc12.src.rpm
Having just looked, it carries a small patch (elftoaout-include.patch) which
modifies the Exec struct, though now that I look at it, as mentioned by Mark,
those should probably be u_int32_t for portability between 32/64.
We didn't make any changes to elftoaout so we never hit this particular problem.
Reply to: