Hello Thorsten Glaser,
Am 24.03.19 um 14:25 schrieb Thorsten Glaser:
> Bernhard Übelacker dixit:
>
>> I see that the syscall number gets modified to become 0x4062.
>>
>> But the syscall modifies 144 bytes, more than just the size of
>> variable ru1 of 88 bytes.
>>
>> This 144 bytes is the siz
Bernhard Übelacker dixit:
>I see that the syscall number gets modified to become 0x4062.
>
>But the syscall modifies 144 bytes, more than just the size of
>variable ru1 of 88 bytes.
>
>This 144 bytes is the size I could observe within amd64 userland.
The x32 syscalls often have struct mapping
Hello Thorsten,
Am 24.03.19 um 01:46 schrieb Thorsten Glaser:
> Bernhard Übelacker dixit:
>
>> Now I wonder if a x32 binary doing a syscall getrusage to a 64bit
>> kernel is supposed to supply memory like "long" would be 8 bytes?
>
> The x32 kernel is amd64, it’s just a userspace ABI, with a
> s
Bernhard Übelacker dixit:
>Now I wonder if a x32 binary doing a syscall getrusage to a 64bit
>kernel is supposed to supply memory like "long" would be 8 bytes?
The x32 kernel is amd64, it’s just a userspace ABI, with a
separate entry point in the kernel (OR’d a high bit in the
syscall number).
l
Dear Maintainer,
I just tried to get some more information.
Following backtrace, little after the stack smashing happend, I could
reproduce with a amd64 qemu VM crossgraded from Buster amd64 to x32. :-)
That "struct rusage" has some elements of type "long" [1].
In gdb sizeof shows 4 bytes for suc
Package: dietlibc-dev
Version: 0.34~cvs20160606-10
Severity: important
https://buildd.debian.org/status/fetch.php?pkg=mksh&arch=x32&ver=57-1&stamp=1551461619&raw=0
mksh built, with -fstack-protector-strong though, against dietlibc is
not usable at all, on x32 only.
-- System Information:
Debian
6 matches
Mail list logo
