Hello :-)
Here are the diff -u. I've used the 5.0-RELEASE sys.tar.gz to compile
the kernel.
dsdt.c : aml_evalexpr now operates with 64-bit unsigned integers as it
should. (quick hack, not throughoutly tested).
--- dsdt.c.old 2012-03-06 10:28:54.000000000 +0100
+++ dsdt.c 2012-03-06 11:04:15.000000000 +0100
@@ -63,7 +63,7 @@
const void *);
u_int64_t aml_convradix(u_int64_t, int, int);
-int64_t aml_evalexpr(int64_t, int64_t, int);
+u_int64_t aml_evalexpr(u_int64_t, u_int64_t, int);
int aml_lsb(u_int64_t);
int aml_msb(u_int64_t);
@@ -1099,10 +1099,10 @@
}
/* Evaluate Math operands */
-int64_t
-aml_evalexpr(int64_t lhs, int64_t rhs, int opcode)
+u_int64_t
+aml_evalexpr(u_int64_t lhs, u_int64_t rhs, int opcode)
{
- int64_t res = 0;
+ u_int64_t res = 0;
switch (opcode) {
/* Math operations */
--------------------
acpiec.c: in acpiec_getcrs: hack to remove the first backslaches of
sc->ec_id to temporary solve the problem with the configuration of
acpiec. (In my case sc->ec=id is prefixed by two backslashes instead of
one or none and this was causing aml_searchname to return NULL).
--- acpiec.c.old 2012-03-06 10:28:54.000000000 +0100
+++ acpiec.c 2012-03-06 11:06:01.000000000 +0100
@@ -394,6 +394,7 @@
int64_t gpe;
struct acpi_ecdt *ecdt = aa->aaa_table;
extern struct aml_node aml_root;
+ u_int8_t *ec_id_temp;
/* Check if this is ECDT initialization */
if (ecdt) {
@@ -407,7 +408,10 @@
ec_data = ecdt->ec_data.address;
/* Get devnode from header */
- sc->sc_devnode = aml_searchname(&aml_root, ecdt->ec_id);
+ ec_id_temp=ecdt->ec_id;
+ while (*ec_id_temp==AMLOP_ROOTCHAR)
+ ec_id_temp++;
+ sc->sc_devnode = aml_searchname(&aml_root, ec_id_temp);
goto ecdtdone;
}
--------
Alternatively, a more general way to remove the superfluous backslash is
to make it directly in aml_searchname but that doesn't seem to make any
difference in my case:
--- dsdt.c.old 2012-03-06 10:28:54.000000000 +0100
+++ dsdt.c 2012-03-06 11:05:01.000000000 +0100
@@ -4085,7 +4085,7 @@
int i;
dnprintf(25,"Searchname: %s:%s = ", aml_nodename(root), vname);
- if (*name == AMLOP_ROOTCHAR) {
+ while (*name == AMLOP_ROOTCHAR) {
root = &aml_root;
name++;
}
To summarize, these fixes have solved the following problems on my EeePC
Asus 1001PX:
64-bit unsigned integers virtual machine:
- hw.sensors.acpitz0.temp0 returns now the correct temperature (or 60B0
if acpiec is not configured) instead of 255B0C! (Ones was considered
smaller than 255)
- hw.sensors.acpibat0.amphour3 (remaining capacity) returns 4Ah (the
full capicity) instead of 65Ah! (I think it's because of this operation
in \SB.PCI0.SBRG.EC0.BST2: Divide ffffffffffffff9c 63 =
ffffffffffffffff).
After fixing the problem with acpiec :
- temperatures reported by _TMP is now precise.
- hw.sensors.acpibat0.volt1 returns 12V instead of 65V!
- hw.sensors.acpibat0.raw0 returns 0 (battery full, correct in my case)
instead of 2 (battery discharging).
- hw.sensors.acpibat0.raw1 returns 0 instead of 1 (I don't know to waht
this corresponds to).
- fan has been activated but I didn't tested that a lot.
Best regards,
Christophe Staiesse.
Le lundi 05 mars 2012 C 20:28 +0000, Alexey E. Suslikov a C)crit :
> Christophe StaC/esse <chastai <at> skynet.be> writes:
>
> > To solve the problem temporarily I've simply changed the types of the
> > arguments and the return type of aml_evalexpr to u_int64_t. And now,
>
> Hey.
>
> Could you provide diff -u so other people can try this out?
>
> > (SB.PCI0.SBRG.EC0) is prefixed by several backslashes characters (4 was
> > displayed in the log but I think it is 2 in reality), what
> > aml_searchname is not supposed to handle apparently and so returns NULL.
> > Keeping one or none backslash solve the problem temporarily at least.
>
> Same as above: diff -u please :)
>
> Alexey
