Create new i386 windows/bsd thunking layer, similar to the amd64 thunking

layer, but with a twist.

The twist has to do with the fact that Microsoft supports structured
exception handling in kernel mode. On the i386 arch, exception handling
is implemented by hanging an exception registration list off the
Thread Environment Block (TEB), and the TEB is accessed via the %fs
register. The problem is, we use %fs as a pointer to the pcpu stucture,
which means any driver that tries to write through %fs:0 will overwrite
the curthread pointer and make a serious mess of things.

To get around this, Project Evil now creates a special entry in
the GDT on each processor. When we call into Windows code, a context
switch routine will fix up %fs so it points to our new descriptor,
which in turn points to a fake TEB. When the Windows code returns,
or calls out to an external routine, we swap %fs back again. Currently,
Project Evil makes use of GDT slot 7, which is all 0s by default.
I fully expect someone to jump up and say I can't do that, but I
couldn't find any code that makes use of this entry anywhere. Sadly,
this was the only method I could come up with that worked on both
UP and SMP. (Modifying the LDT works on UP, but becomes incredibly
complicated on SMP.) If necessary, the context switching stuff can
be yanked out while preserving the convention calling wrappers.

(Fortunately, it looks like Microsoft uses some special epilog/prolog
code on amd64 to implement exception handling, so the same nastiness
won't be necessary on that arch.)

The advantages are:

- Any driver that uses %fs as though it were a TEB pointer won't
  clobber pcpu.
- All the __stdcall/__fastcall/__regparm stuff that's specific to
  gcc goes away.

Also, while I'm here, switch NdisGetSystemUpTime() back to using
nanouptime() again. It turns out nanouptime() is way more accurate
than just using ticks(). On slower machines, the Atheros drivers
I tested seem to take a long time to associate due to the loss
in accuracy.

1 files changed, 26 insertions, 17 deletions

diff --git a/sys/dev/if_ndis/if_ndis.c b/sys/dev/if_ndis/if_ndis.c
index 57dd266..45da8c9 100644
--- a/sys/dev/if_ndis/if_ndis.c
+++ b/sys/dev/if_ndis/if_ndis.c
@@ -94,13 +94,10 @@ void ndis_shutdown		(device_t);
 
 int ndisdrv_modevent		(module_t, int, void *);
 
-static __stdcall void ndis_txeof	(ndis_handle,
-	ndis_packet *, ndis_status);
-static __stdcall void ndis_rxeof	(ndis_handle,
-	ndis_packet **, uint32_t);
-static __stdcall void ndis_linksts	(ndis_handle,
-	ndis_status, void *, uint32_t);
-static __stdcall void ndis_linksts_done	(ndis_handle);
+static void ndis_txeof		(ndis_handle, ndis_packet *, ndis_status);
+static void ndis_rxeof		(ndis_handle, ndis_packet **, uint32_t);
+static void ndis_linksts	(ndis_handle, ndis_status, void *, uint32_t);
+static void ndis_linksts_done	(ndis_handle);
 
 /* We need to wrap these functions for amd64. */
 
@@ -157,12 +154,16 @@ ndisdrv_modevent(mod, cmd, arg)
 		ndisdrv_loaded++;
                 if (ndisdrv_loaded > 1)
 			break;
-		windrv_load(mod, (vm_offset_t)drv_data, 0);
-		windrv_wrap((funcptr)ndis_rxeof, &ndis_rxeof_wrap);
-		windrv_wrap((funcptr)ndis_txeof, &ndis_txeof_wrap);
-		windrv_wrap((funcptr)ndis_linksts, &ndis_linksts_wrap);
+		if (windrv_load(mod, (vm_offset_t)drv_data, 0))
+			return(EINVAL);
+		windrv_wrap((funcptr)ndis_rxeof, &ndis_rxeof_wrap,
+		    3, WINDRV_WRAP_STDCALL);
+		windrv_wrap((funcptr)ndis_txeof, &ndis_txeof_wrap,
+		    3, WINDRV_WRAP_STDCALL);
+		windrv_wrap((funcptr)ndis_linksts, &ndis_linksts_wrap,
+		    4, WINDRV_WRAP_STDCALL);
 		windrv_wrap((funcptr)ndis_linksts_done,
-		    &ndis_linksts_done_wrap);
+		    &ndis_linksts_done_wrap, 1, WINDRV_WRAP_STDCALL);
 		break;
 	case MOD_UNLOAD:
 		ndisdrv_loaded--;
@@ -493,7 +494,15 @@ ndis_attach(dev)
 	 */
 
 	img = drv_data;
+
 	drv = windrv_lookup((vm_offset_t)img, NULL);
+
+	if (drv == NULL) {
+		device_printf(dev, "failed to find driver_object!\n");
+		error = ENXIO;
+		goto fail;
+	}
+
 	if (NdisAddDevice(drv, pdo) != STATUS_SUCCESS) {
 		device_printf(dev, "failed to create FDO!\n");
 		error = ENXIO;
@@ -950,7 +959,7 @@ ndis_resume(dev)
  * copy the packet data into local storage and let the driver keep the
  * packet.
  */
-__stdcall static void
+static void
 ndis_rxeof(adapter, packets, pktcnt)
 	ndis_handle		adapter;
 	ndis_packet		**packets;
@@ -1030,7 +1039,7 @@ ndis_rxeof(adapter, packets, pktcnt)
  * A frame was downloaded to the chip. It's safe for us to clean up
  * the list buffers.
  */
-__stdcall static void
+static void
 ndis_txeof(adapter, packet, status)
 	ndis_handle		adapter;
 	ndis_packet		*packet;
@@ -1072,7 +1081,7 @@ ndis_txeof(adapter, packet, status)
 	return;
 }
 
-__stdcall static void
+static void
 ndis_linksts(adapter, status, sbuf, slen)
 	ndis_handle		adapter;
 	ndis_status		status;
@@ -1090,7 +1099,7 @@ ndis_linksts(adapter, status, sbuf, slen)
 	return;
 }
 
-__stdcall static void
+static void
 ndis_linksts_done(adapter)
 	ndis_handle		adapter;
 {
@@ -1178,7 +1187,7 @@ ndis_ticktask(xsc)
 	void			*xsc;
 {
 	struct ndis_softc	*sc;
-	__stdcall ndis_checkforhang_handler hangfunc;
+	ndis_checkforhang_handler hangfunc;
 	uint8_t			rval;
 	ndis_media_state	linkstate;
 	int			error, len;