Re: [Qemu-devel] [PATCH] linux-user: /proc/self/maps content

2014-08-06 Thread Riku Voipio
On Tue, Aug 05, 2014 at 05:24:27PM +0400, Mikhail Ilin wrote: > I've tested the sample for Aarch64 myself and found that the > approach should also work fine. > > Translation layout: > > $ qemu-aarch64 -strace /tmp/busybox-static cat /proc/self/maps > > startend size

Re: [Qemu-devel] [PATCH] linux-user: /proc/self/maps content

2014-08-05 Thread Mikhail Ilyin
Build /proc/self/maps doing a match against guest memory translation table. Output only that map records which are valid for guest memory layout. Signed-off-by: Mikhail Ilyin --- The previous patch won't compile with 32 bits compiler because of wrong casting type, replace uint64_t with uintptr_t

Re: [Qemu-devel] [PATCH] linux-user: /proc/self/maps content

2014-08-05 Thread Mikhail Ilin
Thanks, I've tested the sample for Aarch64 myself and found that the approach should also work fine. Translation layout: $ qemu-aarch64 -strace /tmp/busybox-static cat /proc/self/maps startend size prot 0040-005ba000 001ba000 r-x

Re: [Qemu-devel] [PATCH] linux-user: /proc/self/maps content

2014-08-05 Thread Riku Voipio
Hi, On Tue, Aug 05, 2014 at 03:10:07PM +0400, Mikhail Ilyin wrote: > Build /proc/self/maps doing a match against guest memory translation table. > Output only that map records which are valid for guest memory layout. This is clear improvement, for most archs. But seems aarch64, openrisc still lea

[Qemu-devel] [PATCH] linux-user: /proc/self/maps content

2014-08-05 Thread Mikhail Ilyin
Build /proc/self/maps doing a match against guest memory translation table. Output only that map records which are valid for guest memory layout. Signed-off-by: Mikhail Ilyin --- include/exec/cpu-all.h | 2 ++ linux-user/syscall.c | 25 ++--- 2 files changed, 12 insertions