The HBA does DMA, thus we must enable the busmaster bit, otherwise dma access will fail with recent qemu versions.
Signed-off-by: Paolo Bonzini pbonzini@redhat.com --- src/esp-scsi.c | 2 ++ 1 file modificato, 2 inserzioni(+)
diff --git a/src/esp-scsi.c b/src/esp-scsi.c index b795012..c43e55b 100644 --- a/src/esp-scsi.c +++ b/src/esp-scsi.c @@ -202,6 +202,8 @@ init_esp_scsi(struct pci_device *pci) pci_bdf_to_bus(bdf), pci_bdf_to_dev(bdf), pci_bdf_to_fn(bdf), iobase);
+ pci_config_maskw(bdf, PCI_COMMAND, 0, PCI_COMMAND_MASTER); + // reset outb(ESP_CMD_RESET, iobase + ESP_CMD);
Il 20/11/2012 18:33, Paolo Bonzini ha scritto:
The HBA does DMA, thus we must enable the busmaster bit, otherwise dma access will fail with recent qemu versions.
Signed-off-by: Paolo Bonzini pbonzini@redhat.com
As usual we forget to update SeaBIOS. We probably want these two from Gerd:
7d05257 (lsi: enable busmaster, 2012-11-20) 455a7c8 (add acpi pmtimer support, 2012-09-06)
and a matching patch to enable bus master DMA in esp.c (just submitted).
The choice is between committing these three patches on top of 1.7.1, or just picking the top of the SeaBIOS tree. There hasn't been much action since 1.7.1 (mostly SeaVGABIOS changes), so I prefer the latter.
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
Paolo
On Tue, Nov 20, 2012 at 06:42:25PM +0100, Paolo Bonzini wrote:
Il 20/11/2012 18:33, Paolo Bonzini ha scritto:
The HBA does DMA, thus we must enable the busmaster bit, otherwise dma access will fail with recent qemu versions.
Signed-off-by: Paolo Bonzini pbonzini@redhat.com
As usual we forget to update SeaBIOS. We probably want these two from Gerd:
7d05257 (lsi: enable busmaster, 2012-11-20) 455a7c8 (add acpi pmtimer support, 2012-09-06)
and a matching patch to enable bus master DMA in esp.c (just submitted).
The choice is between committing these three patches on top of 1.7.1, or just picking the top of the SeaBIOS tree. There hasn't been much action since 1.7.1 (mostly SeaVGABIOS changes), so I prefer the latter.
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
-Kevin
Hi,
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Agree. I'll prepare a seabios update pull once the tag is there.
cheers, Gerd
Hi,
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
cheers, Gerd
On Mon, Jan 07, 2013 at 03:29:33PM +0100, Gerd Hoffmann wrote:
Hi,
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
Lets target Jan 18th for the next SeaBIOS release.
-Kevin
On Mon, Jan 07, 2013 at 06:34:01PM -0500, Kevin O'Connor wrote:
On Mon, Jan 07, 2013 at 03:29:33PM +0100, Gerd Hoffmann wrote:
Hi,
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
Why is 1.4 being rushed so quick? Is this because 1.3.0 was so broken or is there something else?
On Mon, Jan 07, 2013 at 06:51:03PM -0500, Brad Smith wrote:
On Mon, Jan 07, 2013 at 06:34:01PM -0500, Kevin O'Connor wrote:
On Mon, Jan 07, 2013 at 03:29:33PM +0100, Gerd Hoffmann wrote:
Hi,
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
Why is 1.4 being rushed so quick? Is this because 1.3.0 was so broken or is there something else?
I don't understand your questions. The 1.4 release is for QEMU and I don't control its release schedule.
-Kevin
On Mon, Jan 07, 2013 at 06:56:00PM -0500, Kevin O'Connor wrote:
On Mon, Jan 07, 2013 at 06:51:03PM -0500, Brad Smith wrote:
On Mon, Jan 07, 2013 at 06:34:01PM -0500, Kevin O'Connor wrote:
On Mon, Jan 07, 2013 at 03:29:33PM +0100, Gerd Hoffmann wrote:
Hi,
Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these days? Other opinions?
A new release can be produced if desired. Given that there isn't much new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
Why is 1.4 being rushed so quick? Is this because 1.3.0 was so broken or is there something else?
I don't understand your questions. The 1.4 release is for QEMU and I don't control its release schedule.
Of course the 1.4 release is for QEMU, that's pretty obvious. Sorry the question wasn't necessarily for you. I just noticed what was quoted when you had replied.
Il 08/01/2013 00:51, Brad Smith ha scritto:
> >> Kevin, are you okay with tagging SeaBIOS master as 1.7.2 one of these > >> days? Other opinions? > > A new release can be produced if desired. Given that there isn't much > new since 1.7.1 I too would prefer to just push forward to 1.7.2.
Ping? Meanwhile 1.4 freeze is in sight (soft freeze in one week, hard freeze + -rc0 feb 1st).
Why is 1.4 being rushed so quick? Is this because 1.3.0 was so broken or is there something else?
Anthony decides, as release manager.
One reason I recall, is that releases in Mar/Jun/Sep/Dec do not work well with Fedora and Ubuntu freeze dates. Feb/May/Aug/Nov should work a little better.
Paolo
On Tue, Nov 20, 2012 at 06:33:41PM +0100, Paolo Bonzini wrote:
The HBA does DMA, thus we must enable the busmaster bit, otherwise dma access will fail with recent qemu versions.
Thanks - I applied this.
-Kevin
-
Brad Smith -
Gerd Hoffmann -
Kevin O'Connor -
Paolo Bonzini