Re: PATCH: handle non aligned mmaps of sections
Marcus Meissner <marcus(a)jet.franken.de> writes:
The file itself is page aligned in these cases, so don't get bad overlaps.
This is in a codepath which did not trigger before, this is just new with WDM handling.
I don't think it makes sense to use the normal mapping code for such files, we should really have a separate loader for the NT kernel stuff (which can of course re-use some lower level functions of the current loader). -- Alexandre Julliard julliard(a)winehq.com
--- Alexandre Julliard <julliard(a)winehq.org> wrote:
I don't think it makes sense to use the normal mapping code for such files, we should really have a separate loader for the NT kernel stuff (which can of course re-use some lower level functions of the current loader).
Do we also create a stub implementation of ntoskrnl for the driver support? SecDev.sys is reported to load under ReactOS and I would like to help get it to load under WINE. Thanks Steven __________________________________ Do you Yahoo!? Protect your identity with Yahoo! Mail AddressGuard http://antispam.yahoo.com/whatsnewfree
"Steven" == Steven Edwards <steven_ed4153(a)yahoo.com> writes:
Steven> --- Alexandre Julliard <julliard(a)winehq.org> wrote: >> I don't think it makes sense to use the normal mapping code for such >> files, we should really have a separate loader for the NT kernel >> stuff (which can of course re-use some lower level functions of the >> current loader). Steven> Do we also create a stub implementation of ntoskrnl for the Steven> driver support? SecDev.sys is reported to load under ReactOS and Steven> I would like to help get it to load under WINE. More question: How does a user program call into SecDev.sys? Via DeviceIOControl()? I looked at some .sys files, and they don't export anything. How is DeviceIOControl reached? Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------
On Wed, Nov 12, 2003 at 05:59:31PM +0100, Uwe Bonnes wrote:
"Steven" == Steven Edwards <steven_ed4153(a)yahoo.com> writes:
Steven> --- Alexandre Julliard <julliard(a)winehq.org> wrote: >> I don't think it makes sense to use the normal mapping code for such >> files, we should really have a separate loader for the NT kernel >> stuff (which can of course re-use some lower level functions of the >> current loader).
Steven> Do we also create a stub implementation of ntoskrnl for the Steven> driver support? SecDev.sys is reported to load under ReactOS and Steven> I would like to help get it to load under WINE.
More question:
How does a user program call into SecDev.sys? Via DeviceIOControl()? I looked at some .sys files, and they don't export anything. How is DeviceIOControl reached?
On load, the loader calls start(DRIVER_OPEN *do, PUNICODE_STRING *name); The DRIVER_OPEN struct is filled out by the called function, including lots of function pointers. One of them is called from DeviceIoControl(). Ciao, Marcus
On Wed, Nov 12, 2003 at 08:38:03AM -0800, Steven Edwards wrote:
--- Alexandre Julliard <julliard(a)winehq.org> wrote:
I don't think it makes sense to use the normal mapping code for such files, we should really have a separate loader for the NT kernel stuff (which can of course re-use some lower level functions of the current loader).
Do we also create a stub implementation of ntoskrnl for the driver support? SecDev.sys is reported to load under ReactOS and I would like to help get it to load under WINE.
We will need to. I have a very small stub implementation on my laptop, but before fixing the loader it will not make sense to submit it ;) Ciao, Marcus
participants (4)
-
Alexandre Julliard -
Marcus Meissner -
Steven Edwards -
Uwe Bonnes