8:50 a.m.
[ adding linux-nvdimm ]
On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss <erik.schmauss(a)intel.com>
wrote:
ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
This change completes the integration of the recent changes to
package object handling with the module-level code support.
For acpi_exec, the -ep flag is removed.
This change allows table load to behave as if it were a method
invocation. Before this, the definition block definition below would
have loaded all named objects at the root scope. After loading, it
would execute the if statements at the root scope.
DefinitionBlock (...)
{
Name(OBJ1, 0)
if (1)
{
Device (DEV1)
{
Name (_HID,0x0)
}
}
Scope (DEV1)
{
Name (OBJ2)
}
}
The above code would load OBJ1 to the namespace, defer the execution
of the if statement and attempt to add OBJ2 within the scope of DEV1.
Since DEV1 is not in scope, this would incur an AE_NOT_FOUND error.
After this error is emitted, the if block is invoked and DEV1 and its
_HID is added to the namespace.
This commit changes the behavior to execute the if block in place
rather than deferring it until all tables are loaded. The new
behavior is as follows: insert OBJ1 in the namespace, invoke the if
statement and add DEV1 and its _HID to the namespace, add OBJ2 to the
scope of DEV1.
Bug report links:
Link: https://bugs.acpica.org/show_bug.cgi?id=963
Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
ACPICA repo:
Link: https://github.com/acpica/acpica/commit/8faf6fca
Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
Signed-off-by: Bob Moore <robert.moore(a)intel.com>
Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
Hi,
This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
module-level code") regresses the detection of persistent memory in my
qemu-kvm setup.
Expect:
# ndctl list -BR
[
{
"provider":"ACPI.NFIT",
"dev":"ndbus1",
"regions":[
{
"dev":"region1",
"size":33285996544,
"available_size":0,
"type":"pmem",
"numa_node":0,
"iset_id":52512752653219036,
"persistence_domain":"unknown"
},
{
"dev":"region2",
"size":33285996544,
"available_size":0,
"type":"pmem",
"numa_node":0,
"iset_id":52512795602891997,
"persistence_domain":"unknown"
}
]
},
{
"provider":"e820",
"dev":"ndbus0",
"regions":[
{
"dev":"region0",
"size":4294967296,
"available_size":0,
"type":"pmem",
"persistence_domain":"unknown"
}
]
}
]
Actual:
# ndctl list -BR
{
"provider":"e820",
"dev":"ndbus0",
"regions":[
{
"dev":"region0",
"size":4294967296,
"available_size":0,
"type":"pmem",
"persistence_domain":"unknown"
}
]
}
The immediately preceding commit 7decc66df940 ("ACPICA: Revert "Fix for
implicit result conversion for the To____ functions"") works fine.
Commit 5a8361f7ecce does not revert cleanly. Here is my qemu-kvm command
line:
label_size=$((128*1024))
mem_size=$((31 << 30))
kvm=(
$qemu
-enable-kvm
-cpu kvm64
-kernel $kernel
-initrd $initrd
-m 16G,slots=4,maxmem=128G
-machine pc-i440fx-2.4,accel=kvm,usb=off,vmport=off,nvdimm
-cpu SandyBridge
-smp 40
-netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no
-device
virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:b7:a1:ad,bus=pci.0,addr=0x7
-object
memory-backend-file,id=mem1,share,mem-path=${mem}1,size=$((label_size +
mem_size))
-device nvdimm,memdev=mem1,id=nv1,label-size=${label_size}
-object
memory-backend-file,id=mem2,share,mem-path=${mem}2,size=$((label_size +
mem_size))
-device nvdimm,memdev=mem2,id=nv2,label-size=${label_size}
-numa node
-numa node
-device ahci,id=sata0,bus=pci.0,addr=0x8
-drive file=$IMAGE,if=none,id=drive-sata0-0-0,format=raw
-device ide-hd,bus=sata0.0,drive=drive-sata0-0-0,id=sata0-0-0
-boot order=nc
-no-reboot
-watchdog i6300esb
-rtc base=localtime
-serial stdio
-display none
-monitor null
)
...where the important lines that setup the pmem emulation are:
-object
memory-backend-file,id=mem1,share,mem-path=${mem}1,size=$((label_size +
mem_size))
-device nvdimm,memdev=mem1,id=nv1,label-size=${label_size}
-object
memory-backend-file,id=mem2,share,mem-path=${mem}2,size=$((label_size +
mem_size))
-device nvdimm,memdev=mem2,id=nv2,label-size=${label_size}
My qemu-kvm version is: v2.7.0-1059-g63e18bdfd8a0
Let me know if you need any more information or help getting the kvm setup
going.
---
drivers/acpi/acpica/dspkginit.c | 128
++++++++++++++++++++++------------------
drivers/acpi/acpica/dswexec.c | 6 +-
drivers/acpi/acpica/nsparse.c | 5 +-
drivers/acpi/acpica/pstree.c | 1 +
include/acpi/acpixf.h | 8 ++-
5 files changed, 85 insertions(+), 63 deletions(-)
diff --git a/drivers/acpi/acpica/dspkginit.c
b/drivers/acpi/acpica/dspkginit.c
index 902bee78036c..a307a07aeacd 100644
--- a/drivers/acpi/acpica/dspkginit.c
+++ b/drivers/acpi/acpica/dspkginit.c
@@ -47,6 +47,7 @@
#include "amlcode.h"
#include "acdispat.h"
#include "acinterp.h"
+#include "acparser.h"
#define _COMPONENT ACPI_NAMESPACE
ACPI_MODULE_NAME("dspkginit")
@@ -94,12 +95,19 @@ acpi_ds_build_internal_package_obj(struct
acpi_walk_state
*walk_state,
union acpi_parse_object *parent;
union acpi_operand_object *obj_desc = NULL;
acpi_status status = AE_OK;
+ u8 module_level_code = FALSE;
u16 reference_count;
u32 index;
u32 i;
ACPI_FUNCTION_TRACE(ds_build_internal_package_obj);
+ /* Check if we are executing module level code */
+
+ if (walk_state->parse_flags & ACPI_PARSE_MODULE_LEVEL) {
+ module_level_code = TRUE;
+ }
+
/* Find the parent of a possibly nested package */
parent = op->common.parent;
@@ -130,24 +138,44 @@ acpi_ds_build_internal_package_obj(struct
acpi_walk_state
*walk_state,
/*
* Allocate the element array (array of pointers to the individual
- * objects) based on the num_elements parameter. Add an extra
pointer slot
- * so that the list is always null terminated.
+ * objects) if necessary. the count is based on the num_elements
+ * parameter. Add an extra pointer slot so that the list is always
+ * null terminated.
*/
- obj_desc->package.elements = ACPI_ALLOCATE_ZEROED(((acpi_size)
- element_count +
- 1) *
sizeof(void *));
-
if (!obj_desc->package.elements) {
- acpi_ut_delete_object_desc(obj_desc);
- return_ACPI_STATUS(AE_NO_MEMORY);
+ obj_desc->package.elements =
ACPI_ALLOCATE_ZEROED(((acpi_size)
+
element_count
> + +
> + 1) *
+
sizeof(void
+
*));
+
> + if (!obj_desc->package.elements) {
> + acpi_ut_delete_object_desc(obj_desc);
> + return_ACPI_STATUS(AE_NO_MEMORY);
> + }
+
> + obj_desc->package.count =
element_count;
> }
- obj_desc->package.count = element_count;
+ /* First arg is element count. Second arg begins the initializer
list */
+
> arg = op->common.value.arg;
> arg = arg->common.next;
> - if (arg) {
> - obj_desc->package.flags |= AOPOBJ_DATA_VALID;
> + /*
> + * If we are executing module-level code, we will defer the
> + * full resolution of the package elements in order to support
> + * forward references from the elements. This provides
> + * compatibility with other ACPI implementations.
> + */
> + if (module_level_code) {
> + obj_desc->package.aml_start = walk_state->aml;
> + obj_desc->package.aml_length = 0;
+
> + ACPI_DEBUG_PRINT_RAW((ACPI_DB_PARSE,
> + "%s: Deferring resolution of
Package elements\n",
+ ACPI_GET_FUNCTION_NAME));
}
/*
@@ -187,15 +215,19 @@ acpi_ds_build_internal_package_obj(struct
acpi_walk_state
*walk_state,
"****DS namepath
not found"));
}
- /*
- * Initialize this package element. This function
handles
the
- * resolution of named references within the
package.
- */
- acpi_ds_init_package_element(0,
- obj_desc->package.
- elements[i], NULL,
- &obj_desc->package.
- elements[i]);
+ if (!module_level_code) {
+ /*
+ * Initialize this package element. This
function
handles the
+ * resolution of named references
within
the package.
+ * Forward references from
module-level
code are deferred
> + * until all ACPI tables are loaded.
> + */
> + acpi_ds_init_package_element(0,
+
obj_desc->package.
+
elements[i],
NULL,
+
&obj_desc->package.
+
elements[i]);
+ }
}
if (*obj_desc_ptr) {
@@ -265,15 +297,21 @@ acpi_ds_build_internal_package_obj(struct
acpi_walk_state
*walk_state,
* num_elements count.
*
* Note: this is not an error, the package is padded out
- * with NULLs.
+ * with NULLs as per the ACPI specification.
*/
- ACPI_DEBUG_PRINT((ACPI_DB_INFO,
- "Package List length (%u) smaller than
NumElements "
- "count (%u), padded with null
elements\n",
- i, element_count));
+ ACPI_DEBUG_PRINT_RAW((ACPI_DB_INFO,
+ "%s: Package List length (%u)
smaller
than NumElements "
+ "count (%u), padded with
null
elements\n",
> + ACPI_GET_FUNCTION_NAME, i,
> + element_count));
> + }
+
> + /* Module-level packages will be resolved later
*/
+
> + if (!module_level_code) {
> + obj_desc->package.flags |= AOPOBJ_DATA_VALID;
> }
- obj_desc->package.flags |= AOPOBJ_DATA_VALID;
op->common.node = ACPI_CAST_PTR(struct acpi_namespace_node,
obj_desc);
return_ACPI_STATUS(status);
}
@@ -363,6 +401,10 @@ acpi_ds_resolve_package_element(union
acpi_operand_object
**element_ptr)
/* Check if reference element is already resolved */
if (element->reference.resolved) {
+ ACPI_DEBUG_PRINT_RAW((ACPI_DB_PARSE,
+ "%s: Package element is already
resolved\n",
+ ACPI_GET_FUNCTION_NAME));
+
> return_VOID;
> }
@@ -383,7 +425,10 @@ acpi_ds_resolve_package_element(union
acpi_operand_object **element_ptr)
"Could not find/resolve named
package
element: %s",
external_path));
> + /* Not found, set the element to NULL */
+
> ACPI_FREE(external_path);
> + acpi_ut_remove_reference(*element_ptr);
> *element_ptr = NULL;
> return_VOID;
> } else if (resolved_node->type == ACPI_TYPE_ANY) {
> @@ -397,23 +442,6 @@ acpi_ds_resolve_package_element(union
acpi_operand_object **element_ptr)
*element_ptr = NULL;
return_VOID;
}
-#if 0
- else if (resolved_node->flags & ANOBJ_TEMPORARY) {
- /*
- * A temporary node found here indicates that the
reference is
- * to a node that was created within this method. We
are
not
- * going to allow it (especially if the package is
returned
- * from the method) -- the temporary node will be
deleted
out
- * from under the method. (05/2017).
- */
- ACPI_ERROR((AE_INFO,
- "Package element refers to a temporary name
[%4.4s], "
- "inserting a NULL element",
- resolved_node->name.ascii));
- *element_ptr = NULL;
- return_VOID;
- }
-#endif
/*
* Special handling for Alias objects. We need resolved_node to
point
@@ -449,20 +477,6 @@ acpi_ds_resolve_package_element(union
acpi_operand_object **element_ptr)
if (ACPI_FAILURE(status)) {
return_VOID;
}
-#if 0
-/* TBD - alias support */
- /*
- * Special handling for Alias objects. We need to setup the type
- * and the Op->Common.Node to point to the Alias target. Note,
- * Alias has at most one level of indirection internally.
- */
- type = op->common.node->type;
- if (type == ACPI_TYPE_LOCAL_ALIAS) {
- type = obj_desc->common.type;
- op->common.node = ACPI_CAST_PTR(struct
acpi_namespace_node,
-
op->common.node->object);
- }
-#endif
switch (type) {
/*
diff --git a/drivers/acpi/acpica/dswexec.c b/drivers/acpi/acpica/dswexec.c
index 2c07d220a50f..46962e34fc02 100644
--- a/drivers/acpi/acpica/dswexec.c
+++ b/drivers/acpi/acpica/dswexec.c
@@ -576,8 +576,10 @@ acpi_status acpi_ds_exec_end_op(struct
acpi_walk_state
*walk_state)
case AML_TYPE_CREATE_OBJECT:
ACPI_DEBUG_PRINT((ACPI_DB_EXEC,
- "Executing CreateObject
(Buffer/Package) Op=%p AMLPtr=%p\n",
- op, op->named.data));
+ "Executing CreateObject
(Buffer/Package) Op=%p Child=%p ParentOpcode=%4.4X\n",
+ op,
op->named.value.arg,
+ op->common.parent->common.
+ aml_opcode));
switch
(op->common.parent->common.aml_opcode) {
case AML_NAME_OP:
diff --git a/drivers/acpi/acpica/nsparse.c b/drivers/acpi/acpica/nsparse.c
index 6ac2d26a2cfb..acb1aede720e 100644
--- a/drivers/acpi/acpica/nsparse.c
+++ b/drivers/acpi/acpica/nsparse.c
@@ -267,8 +267,9 @@ acpi_ns_parse_table(u32 table_index, struct
acpi_namespace_node
*start_node)
ACPI_FUNCTION_TRACE(ns_parse_table);
if (acpi_gbl_parse_table_as_term_list) {
- ACPI_DEBUG_PRINT((ACPI_DB_PARSE,
- "**** Start table execution pass\n"));
+ ACPI_DEBUG_PRINT_RAW((ACPI_DB_PARSE,
+ "%s: **** Start table execution
pass\n",
+ ACPI_GET_FUNCTION_NAME));
status = acpi_ns_execute_table(table_index,
start_node);
if (ACPI_FAILURE(status)) {
diff --git a/drivers/acpi/acpica/pstree.c b/drivers/acpi/acpica/pstree.c
index f9fa88c79b32..a4dd08eca47c 100644
--- a/drivers/acpi/acpica/pstree.c
+++ b/drivers/acpi/acpica/pstree.c
@@ -295,6 +295,7 @@ union acpi_parse_object *acpi_ps_get_child(union
acpi_parse_object *op)
case AML_BUFFER_OP:
case AML_PACKAGE_OP:
+ case AML_VARIABLE_PACKAGE_OP:
case AML_METHOD_OP:
case AML_IF_OP:
case AML_WHILE_OP:
diff --git a/include/acpi/acpixf.h b/include/acpi/acpixf.h
index c2bf1255f5aa..84c946882589 100644
--- a/include/acpi/acpixf.h
+++ b/include/acpi/acpixf.h
@@ -192,15 +192,19 @@ ACPI_INIT_GLOBAL(u8, acpi_gbl_do_not_use_xsdt,
FALSE);
/*
* Optionally support group module level code.
+ * NOTE, this is essentially obsolete and will be removed soon
+ * (01/2018).
*/
-ACPI_INIT_GLOBAL(u8, acpi_gbl_group_module_level_code, TRUE);
+ACPI_INIT_GLOBAL(u8, acpi_gbl_group_module_level_code, FALSE);
/*
* Optionally support module level code by parsing the entire table as
* a term_list. Default is FALSE, do not execute entire table until some
* lock order issues are fixed.
+ * NOTE, this is essentially obsolete and will be removed soon
+ * (01/2018).
*/
-ACPI_INIT_GLOBAL(u8, acpi_gbl_parse_table_as_term_list, FALSE);
+ACPI_INIT_GLOBAL(u8, acpi_gbl_parse_table_as_term_list, TRUE);
/*
* Optionally use 32-bit FADT addresses if and when there is a conflict
--
2.14.3
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
10:56 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
On Thu, Apr 12, 2018 at 3:50 PM, Dan Williams <dan.j.williams(a)intel.com> wrote:
[ adding linux-nvdimm ]
On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss <erik.schmauss(a)intel.com>
wrote:
> ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
> This change completes the integration of the recent changes to
> package object handling with the module-level code support.
> For acpi_exec, the -ep flag is removed.
> This change allows table load to behave as if it were a method
> invocation. Before this, the definition block definition below would
> have loaded all named objects at the root scope. After loading, it
> would execute the if statements at the root scope.
> DefinitionBlock (...)
> {
> Name(OBJ1, 0)
> if (1)
> {
> Device (DEV1)
> {
> Name (_HID,0x0)
> }
> }
> Scope (DEV1)
> {
> Name (OBJ2)
> }
> }
> The above code would load OBJ1 to the namespace, defer the execution
> of the if statement and attempt to add OBJ2 within the scope of DEV1.
> Since DEV1 is not in scope, this would incur an AE_NOT_FOUND error.
> After this error is emitted, the if block is invoked and DEV1 and its
> _HID is added to the namespace.
> This commit changes the behavior to execute the if block in place
> rather than deferring it until all tables are loaded. The new
> behavior is as follows: insert OBJ1 in the namespace, invoke the if
> statement and add DEV1 and its _HID to the namespace, add OBJ2 to the
> scope of DEV1.
> Bug report links:
> Link: https://bugs.acpica.org/show_bug.cgi?id=963
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
> ACPICA repo:
> Link: https://github.com/acpica/acpica/commit/8faf6fca
> Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
> Signed-off-by: Bob Moore <robert.moore(a)intel.com>
> Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
Hi,
This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
module-level code") regresses the detection of persistent memory in my
qemu-kvm setup.
With the following set of clean reverts on top of latest-Linus I'm
back to a working configuration:
Revert "ACPICA: Change a compile-time option to a runtime option"
Revert "ACPICA: Cleanup/simplify module-level code support"
Revert "ACPICA: Rename a global for clarity, no functional change"
Revert "ACPICA: Add option to disable Package object name resolution errors"
Revert "ACPICA: Integrate package handling with module-level code"
...i.e.
git revert 34f206fd757c
git revert a406dea82af8
git revert e7d970f6fca8
git revert 959c38a7e128
git revert 5a8361f7ecce
2:54 p.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
-----Original Message-----
From: Dan Williams [mailto:dan.j.williams@intel.com]
Sent: Thursday, April 12, 2018 5:57 PM
To: Schmauss, Erik <erik.schmauss(a)intel.com>
Cc: Linux ACPI <linux-acpi(a)vger.kernel.org>; Rafael J. Wysocki
<rjw(a)rjwysocki.net>; Moore, Robert <robert.moore(a)intel.com>; linux-nvdimm
<linux-nvdimm(a)lists.01.org>
Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with module-
level code
On Thu, Apr 12, 2018 at 3:50 PM, Dan Williams <dan.j.williams(a)intel.com>
wrote:
> [ adding linux-nvdimm ]
>
> On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss
> <erik.schmauss(a)intel.com>
> wrote:
>
>> ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
>
>> This change completes the integration of the recent changes to
>> package object handling with the module-level code support.
>
>> For acpi_exec, the -ep flag is removed.
>
>> This change allows table load to behave as if it were a method
>> invocation. Before this, the definition block definition below would
>> have loaded all named objects at the root scope. After loading, it
>> would execute the if statements at the root scope.
>
>> DefinitionBlock (...)
>> {
>> Name(OBJ1, 0)
>
>> if (1)
>> {
>> Device (DEV1)
>> {
>> Name (_HID,0x0)
>> }
>> }
>> Scope (DEV1)
>> {
>> Name (OBJ2)
>> }
>> }
>
>> The above code would load OBJ1 to the namespace, defer the execution
>> of the if statement and attempt to add OBJ2 within the scope of DEV1.
>> Since DEV1 is not in scope, this would incur an AE_NOT_FOUND error.
>> After this error is emitted, the if block is invoked and DEV1 and its
>> _HID is added to the namespace.
>
>> This commit changes the behavior to execute the if block in place
>> rather than deferring it until all tables are loaded. The new
>> behavior is as follows: insert OBJ1 in the namespace, invoke the if
>> statement and add DEV1 and its _HID to the namespace, add OBJ2 to the
>> scope of DEV1.
>
>> Bug report links:
>> Link: https://bugs.acpica.org/show_bug.cgi?id=963
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
>
>> ACPICA repo:
>> Link: https://github.com/acpica/acpica/commit/8faf6fca
>
>> Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
>> Signed-off-by: Bob Moore <robert.moore(a)intel.com>
>> Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
>
> Hi,
>
> This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
> module-level code") regresses the detection of persistent memory in my
> qemu-kvm setup.
>
Hi Dan,
Thanks for figuring this out. Do you think it's possible for you to send me a full
acpidump or some sort hexdump of the QEMU environment?
With the following set of clean reverts on top of latest-Linus
I'm back to a
working configuration:
Revert "ACPICA: Change a compile-time option to a runtime option"
Revert "ACPICA: Cleanup/simplify module-level code support"
Revert "ACPICA: Rename a global for clarity, no functional change"
Revert "ACPICA: Add option to disable Package object name resolution errors"
Revert "ACPICA: Integrate package handling with module-level code"
Rafael, we may want to hold back on the module-level code changes (the patches below) for
rc1. Between this and the strange _TSS issue, it seems like there are a few more things to
resolve before this is ready for kernel upstream. If this is the case, I wonder if we
could change the ACPICA version number so that it is one digit lower or append something
to the end of the version number. This would be helpful in reminding us that it's not
the full release in ACPICA upstream. Either that or remove the ACPICA release all together
for this rc..
Thanks,
Erik
...i.e.
git revert 34f206fd757c
git revert a406dea82af8
git revert e7d970f6fca8
git revert 959c38a7e128
git revert 5a8361f7ecce
5:42 p.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
On Friday, April 13, 2018 6:54:04 AM CEST Schmauss, Erik wrote:
> -----Original Message-----
> From: Dan Williams [mailto:dan.j.williams@intel.com]
> Sent: Thursday, April 12, 2018 5:57 PM
> To: Schmauss, Erik <erik.schmauss(a)intel.com>
> Cc: Linux ACPI <linux-acpi(a)vger.kernel.org>; Rafael J. Wysocki
> <rjw(a)rjwysocki.net>; Moore, Robert <robert.moore(a)intel.com>;
linux-nvdimm
> <linux-nvdimm(a)lists.01.org>
> Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with module-
> level code
>
> On Thu, Apr 12, 2018 at 3:50 PM, Dan Williams <dan.j.williams(a)intel.com>
> wrote:
> > [ adding linux-nvdimm ]
> >
> > On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss
> > <erik.schmauss(a)intel.com>
> > wrote:
> >
> >> ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
> >
> >> This change completes the integration of the recent changes to
> >> package object handling with the module-level code support.
> >
> >> For acpi_exec, the -ep flag is removed.
> >
> >> This change allows table load to behave as if it were a method
> >> invocation. Before this, the definition block definition below would
> >> have loaded all named objects at the root scope. After loading, it
> >> would execute the if statements at the root scope.
> >
> >> DefinitionBlock (...)
> >> {
> >> Name(OBJ1, 0)
> >
> >> if (1)
> >> {
> >> Device (DEV1)
> >> {
> >> Name (_HID,0x0)
> >> }
> >> }
> >> Scope (DEV1)
> >> {
> >> Name (OBJ2)
> >> }
> >> }
> >
> >> The above code would load OBJ1 to the namespace, defer the execution
> >> of the if statement and attempt to add OBJ2 within the scope of DEV1.
> >> Since DEV1 is not in scope, this would incur an AE_NOT_FOUND error.
> >> After this error is emitted, the if block is invoked and DEV1 and its
> >> _HID is added to the namespace.
> >
> >> This commit changes the behavior to execute the if block in place
> >> rather than deferring it until all tables are loaded. The new
> >> behavior is as follows: insert OBJ1 in the namespace, invoke the if
> >> statement and add DEV1 and its _HID to the namespace, add OBJ2 to the
> >> scope of DEV1.
> >
> >> Bug report links:
> >> Link: https://bugs.acpica.org/show_bug.cgi?id=963
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
> >
> >> ACPICA repo:
> >> Link: https://github.com/acpica/acpica/commit/8faf6fca
> >
> >> Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
> >> Signed-off-by: Bob Moore <robert.moore(a)intel.com>
> >> Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
> >
> > Hi,
> >
> > This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
> > module-level code") regresses the detection of persistent memory in my
> > qemu-kvm setup.
> >
>
Hi Dan,
Thanks for figuring this out. Do you think it's possible for you to send me a full
acpidump or some sort hexdump of the QEMU environment?
> With the following set of clean reverts on top of latest-Linus I'm back to a
> working configuration:
>
> Revert "ACPICA: Change a compile-time option to a runtime option"
> Revert "ACPICA: Cleanup/simplify module-level code support"
> Revert "ACPICA: Rename a global for clarity, no functional change"
> Revert "ACPICA: Add option to disable Package object name resolution
errors"
> Revert "ACPICA: Integrate package handling with module-level code"
>
Rafael, we may want to hold back on the module-level code changes
(the patches below) for rc1. Between this and the strange _TSS issue,
it seems like there are a few more things to resolve before this is ready
for kernel upstream.
It looks like you are asking me to queue up reverts as per the Dan's report,
is that correct?
If this is the case, I wonder if we could change the
ACPICA version number so that it is one digit lower or append something
to the end of the version number. This would be helpful in reminding us
that it's not the full release in ACPICA upstream. Either that or remove
the ACPICA release all together for this rc..
I can revert the commit updating the ACPICA version number easily enough. :-)
> ...i.e.
>
> git revert 34f206fd757c
> git revert a406dea82af8
> git revert e7d970f6fca8
> git revert 959c38a7e128
> git revert 5a8361f7ecce
3:50 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
On Fri, Apr 13, 2018 at 12:42 AM, Rafael J. Wysocki <rjw(a)rjwysocki.net> wrote:
On Friday, April 13, 2018 6:54:04 AM CEST Schmauss, Erik wrote:
>
> > -----Original Message-----
> > From: Dan Williams [mailto:dan.j.williams@intel.com]
> > Sent: Thursday, April 12, 2018 5:57 PM
> > To: Schmauss, Erik <erik.schmauss(a)intel.com>
> > Cc: Linux ACPI <linux-acpi(a)vger.kernel.org>; Rafael J. Wysocki
> > <rjw(a)rjwysocki.net>; Moore, Robert <robert.moore(a)intel.com>;
linux-nvdimm
> > <linux-nvdimm(a)lists.01.org>
> > Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with module-
> > level code
> >
> > On Thu, Apr 12, 2018 at 3:50 PM, Dan Williams <dan.j.williams(a)intel.com>
> > wrote:
> > > [ adding linux-nvdimm ]
> > >
> > > On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss
> > > <erik.schmauss(a)intel.com>
> > > wrote:
> > >
> > >> ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
> > >
> > >> This change completes the integration of the recent changes to
> > >> package object handling with the module-level code support.
> > >
> > >> For acpi_exec, the -ep flag is removed.
> > >
> > >> This change allows table load to behave as if it were a method
> > >> invocation. Before this, the definition block definition below would
> > >> have loaded all named objects at the root scope. After loading, it
> > >> would execute the if statements at the root scope.
> > >
> > >> DefinitionBlock (...)
> > >> {
> > >> Name(OBJ1, 0)
> > >
> > >> if (1)
> > >> {
> > >> Device (DEV1)
> > >> {
> > >> Name (_HID,0x0)
> > >> }
> > >> }
> > >> Scope (DEV1)
> > >> {
> > >> Name (OBJ2)
> > >> }
> > >> }
> > >
> > >> The above code would load OBJ1 to the namespace, defer the execution
> > >> of the if statement and attempt to add OBJ2 within the scope of DEV1.
> > >> Since DEV1 is not in scope, this would incur an AE_NOT_FOUND error.
> > >> After this error is emitted, the if block is invoked and DEV1 and its
> > >> _HID is added to the namespace.
> > >
> > >> This commit changes the behavior to execute the if block in place
> > >> rather than deferring it until all tables are loaded. The new
> > >> behavior is as follows: insert OBJ1 in the namespace, invoke the if
> > >> statement and add DEV1 and its _HID to the namespace, add OBJ2 to the
> > >> scope of DEV1.
> > >
> > >> Bug report links:
> > >> Link: https://bugs.acpica.org/show_bug.cgi?id=963
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
> > >
> > >> ACPICA repo:
> > >> Link: https://github.com/acpica/acpica/commit/8faf6fca
> > >
> > >> Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
> > >> Signed-off-by: Bob Moore <robert.moore(a)intel.com>
> > >> Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
> > >
> > > Hi,
> > >
> > > This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
> > > module-level code") regresses the detection of persistent memory in
my
> > > qemu-kvm setup.
> > >
> >
> Hi Dan,
>
> Thanks for figuring this out. Do you think it's possible for you to send me a
full acpidump or some sort hexdump of the QEMU environment?
>
> > With the following set of clean reverts on top of latest-Linus I'm back to
a
> > working configuration:
> >
> > Revert "ACPICA: Change a compile-time option to a runtime option"
> > Revert "ACPICA: Cleanup/simplify module-level code support"
> > Revert "ACPICA: Rename a global for clarity, no functional change"
> > Revert "ACPICA: Add option to disable Package object name resolution
errors"
> > Revert "ACPICA: Integrate package handling with module-level code"
> >
>
> Rafael, we may want to hold back on the module-level code changes
> (the patches below) for rc1. Between this and the strange _TSS issue,
> it seems like there are a few more things to resolve before this is ready
> for kernel upstream.
It looks like you are asking me to queue up reverts as per the Dan's report,
is that correct?
Note, that I'm willing to try any proposed fix patches on top. I just
wanted to clarify that we have a simple fallback position if the debug
starts dragging out.
9:15 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
[ trimming ]
>> Rafael, we may want to hold back on the module-level code
changes
>> (the patches below) for rc1. Between this and the strange _TSS issue,
>> it seems like there are a few more things to resolve before this is
>> ready for kernel upstream.
>
Hi Rafael,
> It looks like you are asking me to queue up reverts as per the
Dan's
> report, is that correct?
This is indeed what I meant last week. However, I've looked into the issue and
Dan's qemu
instance had AML that we no longer support. This is because the ACPICA commit makes
changes to the execution of AML
during table load to match windows AML interpreter behavior so this commit also got rid of
support for executing code
containing forward references (except for package elements).
I've suggested a fix for the firmware in a separate email. So I would say that this
issue is resolved after if Dan can run
his test successfully with the adjusted firmware.
If Dan's test is successful, we don’t need to revert these changes
Erik
Note, that I'm willing to try any proposed fix patches on top. I just wanted to
clarify that we have a simple fallback position if the debug starts dragging out.
9:21 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
On Mon, Apr 16, 2018 at 4:15 PM, Schmauss, Erik <erik.schmauss(a)intel.com> wrote:
[ trimming ]
> >> Rafael, we may want to hold back on the module-level code changes
> >> (the patches below) for rc1. Between this and the strange _TSS issue,
> >> it seems like there are a few more things to resolve before this is
> >> ready for kernel upstream.
> >
Hi Rafael,
> > It looks like you are asking me to queue up reverts as per the Dan's
> > report, is that correct?
This is indeed what I meant last week. However, I've looked into the issue and
Dan's qemu
instance had AML that we no longer support. This is because the ACPICA commit makes
changes to the execution of AML
during table load to match windows AML interpreter behavior so this commit also got rid
of support for executing code
containing forward references (except for package elements).
I've suggested a fix for the firmware in a separate email. So I would say that this
issue is resolved after if Dan can run
his test successfully with the adjusted firmware.
If Dan's test is successful, we don’t need to revert these changes
I'm concerned about other qemu-kvm users that do not upgrade their
hypervisor at the same pace as their guest kernel. Especially for
cloud providers that may be running latest mainline kernel on older
qemu-kvm this will look like a pure kernel regression. Is there a
quick fix we can carry in the kernel to support these forward
references, at least until we know that qemu-kvm is no longer shipping
the broken AML?
10:05 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
-----Original Message-----
From: linux-acpi-owner(a)vger.kernel.org [mailto:linux-acpi-
owner(a)vger.kernel.org] On Behalf Of Dan Williams
Sent: Monday, April 16, 2018 4:22 PM
To: Schmauss, Erik <erik.schmauss(a)intel.com>
Cc: Rafael J. Wysocki <rjw(a)rjwysocki.net>; Linux ACPI <linux-
acpi(a)vger.kernel.org>; Moore, Robert <robert.moore(a)intel.com>; linux-
nvdimm <linux-nvdimm(a)lists.01.org>; Qemu Developers <qemu-
devel(a)nongnu.org>
Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with module-
level code
On Mon, Apr 16, 2018 at 4:15 PM, Schmauss, Erik <erik.schmauss(a)intel.com>
wrote:
> [ trimming ]
>> >> Rafael, we may want to hold back on the module-level code changes
>> >> (the patches below) for rc1. Between this and the strange _TSS
>> >> issue, it seems like there are a few more things to resolve before
>> >> this is ready for kernel upstream.
>> >
> Hi Rafael,
>
>> > It looks like you are asking me to queue up reverts as per the
>> > Dan's report, is that correct?
>
> This is indeed what I meant last week. However, I've looked into the
> issue and Dan's qemu instance had AML that we no longer support. This
> is because the ACPICA commit makes changes to the execution of AML
> during table load to match windows AML interpreter behavior so this commit
also got rid of support for executing code containing forward references (except
for package elements).
>
> I've suggested a fix for the firmware in a separate email. So I would
> say that this issue is resolved after if Dan can run his test successfully with
the
adjusted firmware.
>
> If Dan's test is successful, we don’t need to revert these changes
Hi Dan,
I'm concerned about other qemu-kvm users that do not upgrade
their hypervisor
at the same pace as their guest kernel. Especially for cloud providers that may
be running latest mainline kernel on older qemu-kvm this will look like a pure
kernel regression. Is there a quick fix we can carry in the kernel to support these
forward references, at least until we know that qemu-kvm is no longer shipping
the broken AML?
This is a very good point. Thanks for bringing this up! One option is for them to set
the global variable acpi_gbl_execute_tables_as_methods in include/acpi/acpixf.h to
false.
This will effectively revert the new behavior in the AML interpreter and go back to the
old way.
Since this is a global flag, we could have a command line option for Linux kernel to turn
this
feature on.
Out of curiosity, is this ACPI table somehow customized for your work? I have a
collection
of acpi tables and your ACPI tables are the only ones that have an OperationRegion
called
NRAM. What is the chance that others will be running Linux on the same tables as the one
you sent me?
Erik
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the
body of
a message to majordomo(a)vger.kernel.org More majordomo info at
http://vger.kernel.org/majordomo-info.html
2:06 p.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
On Mon, Apr 16, 2018 at 5:05 PM, Schmauss, Erik <erik.schmauss(a)intel.com> wrote:
> -----Original Message-----
> From: linux-acpi-owner(a)vger.kernel.org [mailto:linux-acpi-
> owner(a)vger.kernel.org] On Behalf Of Dan Williams
> Sent: Monday, April 16, 2018 4:22 PM
> To: Schmauss, Erik <erik.schmauss(a)intel.com>
> Cc: Rafael J. Wysocki <rjw(a)rjwysocki.net>; Linux ACPI <linux-
> acpi(a)vger.kernel.org>; Moore, Robert <robert.moore(a)intel.com>; linux-
> nvdimm <linux-nvdimm(a)lists.01.org>; Qemu Developers <qemu-
> devel(a)nongnu.org>
> Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with module-
> level code
>
> On Mon, Apr 16, 2018 at 4:15 PM, Schmauss, Erik <erik.schmauss(a)intel.com>
> wrote:
> > [ trimming ]
> >> >> Rafael, we may want to hold back on the module-level code changes
> >> >> (the patches below) for rc1. Between this and the strange _TSS
> >> >> issue, it seems like there are a few more things to resolve before
> >> >> this is ready for kernel upstream.
> >> >
> > Hi Rafael,
> >
> >> > It looks like you are asking me to queue up reverts as per the
> >> > Dan's report, is that correct?
> >
> > This is indeed what I meant last week. However, I've looked into the
> > issue and Dan's qemu instance had AML that we no longer support. This
> > is because the ACPICA commit makes changes to the execution of AML
> > during table load to match windows AML interpreter behavior so this commit
> also got rid of support for executing code containing forward references (except
> for package elements).
> >
> > I've suggested a fix for the firmware in a separate email. So I would
> > say that this issue is resolved after if Dan can run his test successfully with
the
> adjusted firmware.
> >
> > If Dan's test is successful, we don’t need to revert these changes
>
Hi Dan,
> I'm concerned about other qemu-kvm users that do not upgrade their hypervisor
> at the same pace as their guest kernel. Especially for cloud providers that may
> be running latest mainline kernel on older qemu-kvm this will look like a pure
> kernel regression. Is there a quick fix we can carry in the kernel to support these
> forward references, at least until we know that qemu-kvm is no longer shipping
> the broken AML?
This is a very good point. Thanks for bringing this up! One option is for them to set
the global variable acpi_gbl_execute_tables_as_methods in include/acpi/acpixf.h to
false.
This will effectively revert the new behavior in the AML interpreter and go back to the
old way.
Since this is a global flag, we could have a command line option for Linux kernel to turn
this
feature on.
Out of curiosity, is this ACPI table somehow customized for your work? I have a
collection
of acpi tables and your ACPI tables are the only ones that have an OperationRegion
called
NRAM. What is the chance that others will be running Linux on the same tables as the one
you sent me?
I don't think there's anything atypical about my particular setup. It
creates two virtual NVDIMMs that each represent a 30GB address space.
I suspect any user of the KVM NVDIMM virtualization would see the same
problem.
5:11 a.m.
New subject: [PATCH v2 5/7] ACPICA: Integrate package handling with module-level code
-----Original Message-----
From: Rafael J. Wysocki [mailto:rjw@rjwysocki.net]
Sent: Friday, April 13, 2018 12:43 AM
To: Schmauss, Erik <erik.schmauss(a)intel.com>
Cc: Williams, Dan J <dan.j.williams(a)intel.com>; Linux ACPI <linux-
acpi(a)vger.kernel.org>; Moore, Robert <robert.moore(a)intel.com>; linux-
nvdimm <linux-nvdimm(a)lists.01.org>
Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with
module-level code
On Friday, April 13, 2018 6:54:04 AM CEST Schmauss, Erik wrote:
>
> > -----Original Message-----
> > From: Dan Williams [mailto:dan.j.williams@intel.com]
> > Sent: Thursday, April 12, 2018 5:57 PM
> > To: Schmauss, Erik <erik.schmauss(a)intel.com>
> > Cc: Linux ACPI <linux-acpi(a)vger.kernel.org>; Rafael J. Wysocki
> > <rjw(a)rjwysocki.net>; Moore, Robert <robert.moore(a)intel.com>;
> > linux-nvdimm <linux-nvdimm(a)lists.01.org>
> > Subject: Re: [PATCH v2 5/7] ACPICA: Integrate package handling with
> > module- level code
> >
> > On Thu, Apr 12, 2018 at 3:50 PM, Dan Williams
> > <dan.j.williams(a)intel.com>
> > wrote:
> > > [ adding linux-nvdimm ]
> > >
> > > On Fri, Feb 16, 2018 at 7:20 AM Erik Schmauss
> > > <erik.schmauss(a)intel.com>
> > > wrote:
> > >
> > >> ACPICA commit 8faf6fca445eb7219963d80543fb802302a7a8c7
> > >
> > >> This change completes the integration of the recent changes to
> > >> package object handling with the module-level code support.
> > >
> > >> For acpi_exec, the -ep flag is removed.
> > >
> > >> This change allows table load to behave as if it were a method
> > >> invocation. Before this, the definition block definition below
> > >> would have loaded all named objects at the root scope. After
> > >> loading, it would execute the if statements at the root scope.
> > >
> > >> DefinitionBlock (...)
> > >> {
> > >> Name(OBJ1, 0)
> > >
> > >> if (1)
> > >> {
> > >> Device (DEV1)
> > >> {
> > >> Name (_HID,0x0)
> > >> }
> > >> }
> > >> Scope (DEV1)
> > >> {
> > >> Name (OBJ2)
> > >> }
> > >> }
> > >
> > >> The above code would load OBJ1 to the namespace, defer the
> > >> execution of the if statement and attempt to add OBJ2 within the
scope of DEV1.
> > >> Since DEV1 is not in scope, this would incur an AE_NOT_FOUND
error.
> > >> After this error is emitted, the if block is invoked and DEV1 and
> > >> its _HID is added to the namespace.
> > >
> > >> This commit changes the behavior to execute the if block in place
> > >> rather than deferring it until all tables are loaded. The new
> > >> behavior is as follows: insert OBJ1 in the namespace, invoke the
> > >> if statement and add DEV1 and its _HID to the namespace, add OBJ2
> > >> to the scope of DEV1.
> > >
> > >> Bug report links:
> > >> Link: https://bugs.acpica.org/show_bug.cgi?id=963
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=153541
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=196165
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=192621
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=197207
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198051
> > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=198515
> > >
> > >> ACPICA repo:
> > >> Link: https://github.com/acpica/acpica/commit/8faf6fca
> > >
> > >> Tested-by: Kai-Heng Feng <kai.heng.feng(a)canonical.com>
> > >> Signed-off-by: Bob Moore <robert.moore(a)intel.com>
> > >> Signed-off-by: Erik Schmauss <erik.schmauss(a)intel.com>
> > >
> > > Hi,
> > >
> > > This commit 5a8361f7ecce ("ACPICA: Integrate package handling with
> > > module-level code") regresses the detection of persistent memory
> > > in my qemu-kvm setup.
> > >
> >
> Hi Dan,
>
> Thanks for figuring this out. Do you think it's possible for you to
send me a full acpidump or some sort hexdump of the QEMU environment?
>
> > With the following set of clean reverts on top of latest-Linus I'm
> > back to a working configuration:
> >
> > Revert "ACPICA: Change a compile-time option to a runtime option"
> > Revert "ACPICA: Cleanup/simplify module-level code support"
> > Revert "ACPICA: Rename a global for clarity, no functional change"
> > Revert "ACPICA: Add option to disable Package object name
resolution errors"
> > Revert "ACPICA: Integrate package handling with module-level code"
> >
>
> Rafael, we may want to hold back on the module-level code changes (the
> patches below) for rc1. Between this and the strange _TSS issue, it
> seems like there are a few more things to resolve before this is ready
> for kernel upstream.
It looks like you are asking me to queue up reverts as per the Dan's
report, is that correct?
> If this is the case, I wonder if we could change the ACPICA version
> number so that it is one digit lower or append something to the end of
> the version number. This would be helpful in reminding us that it's
> not the full release in ACPICA upstream. Either that or remove the
> ACPICA release all together for this rc..
I can revert the commit updating the ACPICA version number easily
enough. :-)
[Moore, Robert]
We should be able to figure this one out soon. Need acpidump
Bob
> > ...i.e.
> >
> > git revert 34f206fd757c
> > git revert a406dea82af8
> > git revert e7d970f6fca8
> > git revert 959c38a7e128
> > git revert 5a8361f7ecce
>
1615
days inactive
1620
days old
9 comments
4 participants
participants (4)
-
Dan Williams -
Moore, Robert -
Rafael J. Wysocki -
Schmauss, Erik