Re: ntdll: Stop when we can't map the shared user data

[Vitaliy Margolen]

On 01/14/2011 01:49 PM, André Hentschel wrote:
> It can happen that the adress 0x7ffe0000 is blocked by some loaded libraries already.
> - if (status)
> - {
> -
>    MESSAGE( "wine: failed to map the shared user data: %x\n", status );
> -
>    exit(1);
> - }

Most well behaved programs don't need it. However most copy-protection 
systems do use this area. As well as badly behaved programs (mostly of MS 
origin).

If you running on different architecture then x86 it can be anywhere, since 
you won't be running native windows programs anyway. However some king of 
checking would be nice for all platforms. But don't think it warrants an exit.

Vitaliy.

ntdll: Stop when we can't map the shared user data

[André Hentschel]

Am 16.01.2011 04:45, schrieb Vitaliy Margolen:
> On 01/14/2011 01:49 PM, André Hentschel wrote:
> > It can happen that the adress 0x7ffe0000 is blocked by some loaded libraries already.
> > - if (status)
> > - {
> > -
> >    MESSAGE( "wine: failed to map the shared user data: %x\n", status );
> > -
> >    exit(1);
> > - }
>
> Most well behaved programs don't need it. However most copy-protection systems do use this area. As well as badly behaved programs (mostly of MS origin).
>
> If you running on different architecture then x86 it can be anywhere, since you won't be running native windows programs anyway. However some king of checking would be nice for all platforms. But don't think it warrants an exit.
>
> Vitaliy.

I should have mentioned that it leads to a crash, because later in the function we write to that address.
But you are still right, if it's not needed for normal apps we shouldn't exit. On the other hand it's tricky as we also write to that address from kernel32...