| Message ID | 1373070978-11966-2-git-send-email-agraf@suse.de |
|---|---|
| State | New |
| Headers | show |
On 6 July 2013 01:36, Alexander Graf <agraf@suse.de> wrote: > When forwarding a segmentation fault into the guest process, we were passing > the host's address directly into the guest process's signal descriptor. > > That obviously confused the guest process, since it didn't know what to make > of the (usually 32-bit truncated) address. Passing in h2g(address) makes the > guest process a lot happier. Commit message says it uses h2g, code doesn't. Maybe we should have an h2g_unchecked() for this sort of use? thanks -- PMM
diff --git a/user-exec.c b/user-exec.c index 26cde7c..718c54f 100644 --- a/user-exec.c +++ b/user-exec.c @@ -94,6 +94,12 @@ static inline int handle_cpu_signal(uintptr_t pc, unsigned long address, return 1; } + if (GUEST_BASE) { + /* Convert forcefully to guest address space, invalid addresses + are still valid segv ones */ + address = address - GUEST_BASE; + } + env = current_cpu->env_ptr; /* see if it is an MMU fault */ ret = cpu_handle_mmu_fault(env, address, is_write, MMU_USER_IDX);
When forwarding a segmentation fault into the guest process, we were passing the host's address directly into the guest process's signal descriptor. That obviously confused the guest process, since it didn't know what to make of the (usually 32-bit truncated) address. Passing in h2g(address) makes the guest process a lot happier. This fixes java running in arm-linux-user for me. Signed-off-by: Alexander Graf <agraf@suse.de> --- user-exec.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)