Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de --- src/fw/dsdt_parser.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/src/fw/dsdt_parser.c b/src/fw/dsdt_parser.c index eb5496f3..2ac82821 100644 --- a/src/fw/dsdt_parser.c +++ b/src/fw/dsdt_parser.c @@ -417,6 +417,10 @@ static int parse_termobj(struct parse_state *s, break; case 0x01: /* one */ break; + case 0x06: /* AliasOp */ + offset += parse_namestring(s, ptr + offset, "SourceObject"); + offset += parse_namestring(s, ptr + offset, "AliasObject"); + break; case 0x08: /* name op */ offset += parse_namestring(s, ptr + offset, "name"); offset += parse_termobj(s, ptr + offset);
On Fri, Nov 18, 2022 at 03:27:55PM +0100, Igor Mammedov wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de
Reviewed-by: Gerd Hoffmann kraxel@redhat.com
On Mon, Nov 21, 2022 at 11:12:55AM +0100, Gerd Hoffmann wrote:
On Fri, Nov 18, 2022 at 03:27:55PM +0100, Igor Mammedov wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de
Reviewed-by: Gerd Hoffmann kraxel@redhat.com
Gerd do you think this is doable for this QEMU release? Or do we need a workaround in QEMU?
On Mon, 21 Nov 2022 05:22:40 -0500 "Michael S. Tsirkin" mst@redhat.com wrote:
On Mon, Nov 21, 2022 at 11:12:55AM +0100, Gerd Hoffmann wrote:
On Fri, Nov 18, 2022 at 03:27:55PM +0100, Igor Mammedov wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de
Reviewed-by: Gerd Hoffmann kraxel@redhat.com
Gerd do you think this is doable for this QEMU release? Or do we need a workaround in QEMU?
I saw that there was another fix which you wanted in qemu-7.2, perhaps having to fixes pending it might justify unplanned release to go with qemu-7.2.
PS: I have workaround almost ready for posting, but if we can ship fixed SeaBIOS it would be better.
On Mon, Nov 21, 2022 at 11:12:55AM +0100, Gerd Hoffmann wrote:
On Fri, Nov 18, 2022 at 03:27:55PM +0100, Igor Mammedov wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de
Reviewed-by: Gerd Hoffmann kraxel@redhat.com
Thanks. Looks fine to me as well.
Gerd - feel free to commit this (and the virtio changes as well) when you feel they are ready.
FWIW, if we want a release in the short term, I'd suggest making a v1.16.1 "stable" release.
-Kevin
On Tue, Nov 22, 2022 at 01:32:20PM -0500, Kevin O'Connor wrote:
On Mon, Nov 21, 2022 at 11:12:55AM +0100, Gerd Hoffmann wrote:
On Fri, Nov 18, 2022 at 03:27:55PM +0100, Igor Mammedov wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML) SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard present
due to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedov imammedo@redhat.com Reported-by: Volker Rümelin vr_qemu@t-online.de
Reviewed-by: Gerd Hoffmann kraxel@redhat.com
Thanks. Looks fine to me as well.
Gerd - feel free to commit this (and the virtio changes as well) when you feel they are ready.
Done.
FWIW, if we want a release in the short term, I'd suggest making a v1.16.1 "stable" release.
Well. Currently we have ...
# git log --oneline rel-1.16.0.. 3208b098f51a virtio: finalize features before using device 5ea5c64c20e9 virtio-mmio: read/write the hi 32 features for mmio 61e901bbaadf acpi: parse Alias object 85d56f812f4d virtio-blk: Fix incorrect type conversion in virtio_blk_op() 46de2eec93bf virtio-blk: use larger default request size dc88f9b72df5 malloc: use large ZoneHigh when there is enough memory 3b91e8e9fe93 malloc: use variable for ZoneHigh size 01774004c7f7 reset: force standard PCI configuration access d24f42b0d819 pci: refactor the pci_config_*() functions
... where most changes (all except the two 'malloc' patches) are bugfixes. Cherry-picking 7/9 patches into 1.16.1 feels somewhat silly. Just grabbing a master branch snapshot looks like the simpler solution here.
Independent from that I think it makes sense to plan a new 1.17.0 release.
take care, Gerd
On Wed, Nov 23, 2022 at 11:11:03AM +0100, Gerd Hoffmann wrote:
On Tue, Nov 22, 2022 at 01:32:20PM -0500, Kevin O'Connor wrote:
FWIW, if we want a release in the short term, I'd suggest making a v1.16.1 "stable" release.
Well. Currently we have ...
# git log --oneline rel-1.16.0.. 3208b098f51a virtio: finalize features before using device 5ea5c64c20e9 virtio-mmio: read/write the hi 32 features for mmio 61e901bbaadf acpi: parse Alias object 85d56f812f4d virtio-blk: Fix incorrect type conversion in virtio_blk_op() 46de2eec93bf virtio-blk: use larger default request size dc88f9b72df5 malloc: use large ZoneHigh when there is enough memory 3b91e8e9fe93 malloc: use variable for ZoneHigh size 01774004c7f7 reset: force standard PCI configuration access d24f42b0d819 pci: refactor the pci_config_*() functions... where most changes (all except the two 'malloc' patches) are bugfixes. Cherry-picking 7/9 patches into 1.16.1 feels somewhat silly. Just grabbing a master branch snapshot looks like the simpler solution here.
Well, FWIW, I don't see an issue with tagging the master branch as v1.16.1 .
Independent from that I think it makes sense to plan a new 1.17.0 release.
I'm also fine with making a v1.17.0 release. I'd prefer to schedule that though and I'm not sure the timeframe would meet the qemu release goals.
Cheers, -Kevin
On Wed, Nov 23, 2022 at 11:52:29AM -0500, Kevin O'Connor wrote:
On Wed, Nov 23, 2022 at 11:11:03AM +0100, Gerd Hoffmann wrote:
Well. Currently we have ...
# git log --oneline rel-1.16.0.. 3208b098f51a virtio: finalize features before using device 5ea5c64c20e9 virtio-mmio: read/write the hi 32 features for mmio 61e901bbaadf acpi: parse Alias object 85d56f812f4d virtio-blk: Fix incorrect type conversion in virtio_blk_op() 46de2eec93bf virtio-blk: use larger default request size dc88f9b72df5 malloc: use large ZoneHigh when there is enough memory 3b91e8e9fe93 malloc: use variable for ZoneHigh size 01774004c7f7 reset: force standard PCI configuration access d24f42b0d819 pci: refactor the pci_config_*() functions... where most changes (all except the two 'malloc' patches) are bugfixes. Cherry-picking 7/9 patches into 1.16.1 feels somewhat silly. Just grabbing a master branch snapshot looks like the simpler solution here.
Well, FWIW, I don't see an issue with tagging the master branch as v1.16.1 .
Ok, lets do that then. I'll go tag 1.16.1 next monday unless someone raises objections. Next qemu -rc is planned for tuesday.
take care, Gerd
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML)
The commit dbce582a119 doesn't exist in the QEMU repository. It's commit 47a373faa6.
With best regards, Volker
SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard presentdue to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedovimammedo@redhat.com Reported-by: Volker Rümelinvr_qemu@t-online.de
src/fw/dsdt_parser.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/src/fw/dsdt_parser.c b/src/fw/dsdt_parser.c index eb5496f3..2ac82821 100644 --- a/src/fw/dsdt_parser.c +++ b/src/fw/dsdt_parser.c @@ -417,6 +417,10 @@ static int parse_termobj(struct parse_state *s, break; case 0x01: /* one */ break;
- case 0x06: /* AliasOp */
offset += parse_namestring(s, ptr + offset, "SourceObject");offset += parse_namestring(s, ptr + offset, "AliasObject");break; case 0x08: /* name op */ offset += parse_namestring(s, ptr + offset, "name"); offset += parse_termobj(s, ptr + offset);
On Tue, 22 Nov 2022 20:05:38 +0100 Volker Rümelin vr_qemu@t-online.de wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML)
The commit dbce582a119 doesn't exist in the QEMU repository. It's commit 47a373faa6.
Right, I must have grabbed id from another branch
Gerd, Can you fix up commit message when applying or should I respin patch?
With best regards, Volker
SeaBIOS fails to parse ISA bridge AML with:
parse_termlist: parse error, skip from 92/517 ... ACPI: no PS/2 keyboard presentdue to Alias term in DSDT which isn't handled by SeaBIOS properly. Add dumb Alias parsing which just skips over term, so the rest of AML could be parsed successfully.
Signed-off-by: Igor Mammedovimammedo@redhat.com Reported-by: Volker Rümelinvr_qemu@t-online.de
src/fw/dsdt_parser.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/src/fw/dsdt_parser.c b/src/fw/dsdt_parser.c index eb5496f3..2ac82821 100644 --- a/src/fw/dsdt_parser.c +++ b/src/fw/dsdt_parser.c @@ -417,6 +417,10 @@ static int parse_termobj(struct parse_state *s, break; case 0x01: /* one */ break;
- case 0x06: /* AliasOp */
offset += parse_namestring(s, ptr + offset, "SourceObject");offset += parse_namestring(s, ptr + offset, "AliasObject");break; case 0x08: /* name op */ offset += parse_namestring(s, ptr + offset, "name"); offset += parse_termobj(s, ptr + offset);
On Thu, Nov 24, 2022 at 01:51:44PM +0100, Igor Mammedov wrote:
On Tue, 22 Nov 2022 20:05:38 +0100 Volker Rümelin vr_qemu@t-online.de wrote:
Since QEMU commit dbce582a119 (acpi: pc/q35: drop ad-hoc PCI-ISA bridge AML routines and let bus ennumeration generate AML)
The commit dbce582a119 doesn't exist in the QEMU repository. It's commit 47a373faa6.
Right, I must have grabbed id from another branch
Gerd, Can you fix up commit message when applying or should I respin patch?
It's already committed (with id fixed).
take care, Gerd
-
Gerd Hoffmann -
Igor Mammedov -
Kevin O'Connor -
Michael S. Tsirkin -
Volker Rümelin