Regression in LOAD segment vaddr handling

[SchumBlubBlub]

https://github.com/radare/radare2-regressions/pull/691

$ rabin2 -S bin | grep LOAD0
idx=28 vaddr=0x00400000 paddr=0x00000000 sz=8796 vsz=8796 perm=m-r-x name=LOAD0
$ r2 -qc 'iS' -d bin | grep LOAD0
idx=28 vaddr=0x00000000 paddr=0x00000000 sz=8796 vsz=8796 perm=m-r-x name=LOAD0

lose the vaddr for the segment and r2 iS doesnt meet rabin2 out.

broken by 4fd1e99

[SchumBlubBlub]

wait, this is even worse. It's not just this mangled elf that's the problem. I can't db main on a "hello world" program anymore. 🔥 .

[radare]

dont use db main when you can dcu main. also it works fine for me

…

On 03 Feb 2017, at 14:31, SchumBlubBlub ***@***.***> wrote: wait, this is even worse. It's not just this mangled elf that's the problem. I can't db main on a "hello world" program anymore. 🔥 . — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#6650 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA3-lipFc71nsLJTcjmp70KMDtjvkNdzks5rYywpgaJpZM4L2QZR>.

[leberus]

my fault, i'll check it out now

[Maijin]

There is also e dbg.bep variable that you can set to main

[SchumBlubBlub]

I tried dcu a while ago, its was really slow getting out of the loader and into main, I had to break.

[leberus]

I've found it:

+#define R_IO_PRIV (1 « 5)
+#define R_IO_SHAR (1 « 6)

This breaks it. I'll check how to make that it doesn't collide

[SchumBlubBlub]

There is also e dbg.bep variable that you can set to main

@Maijin , I did, that's how I met this bug.

[leberus]

I think this can be closed. It was solved by 6672, wasn't it?

[radare]

do we have tests?

[leberus]

I promise i'll write some tests for the debugger

[Maijin]

It has tests see the tags. https://github.com/SchumBlubBlub/radare2-regressions/blob/master/t.archos/Linux-x86_64/dbg_seginfo

[SchumBlubBlub]

Tests: link in OP, Maijin just merged: https://github.com/radare/radare2-regressions/pull/691