[rust-dev] Curious about instruction pointer being used to compute string offset

[rust-dev] Curious about instruction pointer being used to compute string offset

[C K Kashyap]

[-- Attachment #1: Type: text/plain, Size: 538 bytes --]

Hi,

Looks like on my ubuntu 64 bit, when I compile

   hello("ABCD");

I get

    719e:       48 8d 05 e0 68 04 00    lea    0x468e0(%rip),%rax        #
4da85 <str1405>
    71a5:       48 89 44 24 08             mov    %rax,0x8(%rsp)
    71aa:       48 c7 44 24 10 04 00    movq   $0x4,0x10(%rsp)
    71b1:       00 00
    71b3:       e8 88 f6 ff ff                 callq  6840
<_ZN5hello20h8ed5f876da0c3862eaaE>


I was wondering why is %rip being used for getting to the string? Or am I
understanding it incorrectly?

Regards,
Kashyap

[-- Attachment #2: Type: text/html, Size: 864 bytes --]

Re: [rust-dev] Curious about instruction pointer being used to compute string offset

[Daniel Micay]

[-- Attachment #1: Type: text/plain, Size: 772 bytes --]

On 09/12/14 10:43 PM, C K Kashyap wrote:
> Hi,
> 
> Looks like on my ubuntu 64 bit, when I compile 
> 
>    hello("ABCD");
> 
> I get 
> 
>     719e:       48 8d 05 e0 68 04 00    lea    0x468e0(%rip),%rax      
>  # 4da85 <str1405>
>     71a5:       48 89 44 24 08             mov    %rax,0x8(%rsp)
>     71aa:       48 c7 44 24 10 04 00    movq   $0x4,0x10(%rsp)
>     71b1:       00 00
>     71b3:       e8 88 f6 ff ff                 callq  6840
> <_ZN5hello20h8ed5f876da0c3862eaaE>
> 
> 
> I was wondering why is %rip being used for getting to the string? Or am
> I understanding it incorrectly?
> 
> Regards,
> Kashyap

That's what position independent code looks like on x86_64.

https://en.wikipedia.org/wiki/Position-independent_code


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [rust-dev] Curious about instruction pointer being used to compute string offset

[C K Kashyap]

[-- Attachment #1: Type: text/plain, Size: 1460 bytes --]

oh nice ... In my kernel I am running into an issue where I am looping
though an &str but it does not get the right characters (all zeros
actually) - https://github.com/ckkashyap/unix/blob/master/kernel/uart.rs

While I've used the PIC option several times, I had not realized that rip
relative addressing is how it's done :) Thanks for the pointer.

Now I need to find out what the problem is in my kernel code ...

Thanks Daniel.

Regards,
Kashyap


On Wed, Dec 10, 2014 at 9:21 AM, Daniel Micay <danielmicay@gmail.com> wrote:

> On 09/12/14 10:43 PM, C K Kashyap wrote:
> > Hi,
> >
> > Looks like on my ubuntu 64 bit, when I compile
> >
> >    hello("ABCD");
> >
> > I get
> >
> >     719e:       48 8d 05 e0 68 04 00    lea    0x468e0(%rip),%rax
> >  # 4da85 <str1405>
> >     71a5:       48 89 44 24 08             mov    %rax,0x8(%rsp)
> >     71aa:       48 c7 44 24 10 04 00    movq   $0x4,0x10(%rsp)
> >     71b1:       00 00
> >     71b3:       e8 88 f6 ff ff                 callq  6840
> > <_ZN5hello20h8ed5f876da0c3862eaaE>
> >
> >
> > I was wondering why is %rip being used for getting to the string? Or am
> > I understanding it incorrectly?
> >
> > Regards,
> > Kashyap
>
> That's what position independent code looks like on x86_64.
>
> https://en.wikipedia.org/wiki/Position-independent_code
>
>
> _______________________________________________
> Rust-dev mailing list
> Rust-dev@mozilla.org
> https://mail.mozilla.org/listinfo/rust-dev
>
>

[-- Attachment #2: Type: text/html, Size: 2471 bytes --]