| Message ID | 1312903262-8212-1-git-send-email-levinsasha928@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
* Sasha Levin <levinsasha928@gmail.com> wrote: > This patch makes BAR 1 16k, instead of BAR0 - which is the PIO bar. > > Signed-off-by: Sasha Levin <levinsasha928@gmail.com> > --- > tools/kvm/hw/vesa.c | 2 +- > tools/kvm/pci.c | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) This changelog is missing some key information: - how did you find the bug (by chance via code review or did you see some actual badness?) - what practical effect (if any) did you see from this patch? - what practical effect (if any) do you expect others to see from this patch? I suspect this patch is only for completeness/correctness but has no practical effect - but that's a guess. Thanks, Ingo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/09/2011 06:39 PM, Ingo Molnar wrote: > * Sasha Levin<levinsasha928@gmail.com> wrote: > > > This patch makes BAR 1 16k, instead of BAR0 - which is the PIO bar. > > > > This changelog is missing some key information: > > - how did you find the bug (by chance via code review or did you see > some actual badness?) > > - what practical effect (if any) did you see from this patch? > > - what practical effect (if any) do you expect others to see from this patch? > > I suspect this patch is only for completeness/correctness but has no > practical effect - but that's a guess. > My guess would be that seabios tried to lay out the BARs and had trouble fitting a 16k pio bar in the small PCI pio region.
On 8/9/11 7:46 PM, Avi Kivity wrote: > On 08/09/2011 06:39 PM, Ingo Molnar wrote: >> * Sasha Levin<levinsasha928@gmail.com> wrote: >> >> > This patch makes BAR 1 16k, instead of BAR0 - which is the PIO bar. >> > >> >> This changelog is missing some key information: >> >> - how did you find the bug (by chance via code review or did you see >> some actual badness?) >> >> - what practical effect (if any) did you see from this patch? >> >> - what practical effect (if any) do you expect others to see from >> this patch? >> >> I suspect this patch is only for completeness/correctness but has no >> practical effect - but that's a guess. >> > > My guess would be that seabios tried to lay out the BARs and had > trouble fitting a 16k pio bar in the small PCI pio region. Sasha? IIRC this fixed some issue with David's out-of-tree PCI driver? -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
I don't know if there were any other drivers for this patch, but it along with another patch (maybe integrated elsewhere for 32bit BAR vs 8bit) certainly helped me out a lot. These patches fixed ioremap errors I was seeing (I had a 16MB PCI memory region, but it appeared to be only 256 bytes in size; the kernel complained bitterly about that on ioremap). It also was an issue of expected vs unexpected output from lspci -vvv. I'm working on my out-of-tree PCI driver to see if it can become in tree. I have more cleanup to do, and seeing how close I can come to the target coding standards. \dae On Wed, Aug 10, 2011 at 09:13:35AM +0300, Pekka Enberg wrote: > On 8/9/11 7:46 PM, Avi Kivity wrote: > >On 08/09/2011 06:39 PM, Ingo Molnar wrote: > >>* Sasha Levin<levinsasha928@gmail.com> wrote: > >> > >>> This patch makes BAR 1 16k, instead of BAR0 - which is the PIO bar. > >>> > >> > >>This changelog is missing some key information: > >> > >> - how did you find the bug (by chance via code review or did you see > >> some actual badness?) > >> > >> - what practical effect (if any) did you see from this patch? > >> > >> - what practical effect (if any) do you expect others to see > >>from this patch? > >> > >>I suspect this patch is only for completeness/correctness but has no > >>practical effect - but that's a guess. > >> > > > >My guess would be that seabios tried to lay out the BARs and had > >trouble fitting a 16k pio bar in the small PCI pio region. > > Sasha? IIRC this fixed some issue with David's out-of-tree PCI driver? -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/tools/kvm/hw/vesa.c b/tools/kvm/hw/vesa.c index 2af08df..9caa6c4 100644 --- a/tools/kvm/hw/vesa.c +++ b/tools/kvm/hw/vesa.c @@ -39,6 +39,7 @@ static struct pci_device_header vesa_pci_device = { .subsys_vendor_id = PCI_SUBSYSTEM_VENDOR_ID_REDHAT_QUMRANET, .subsys_id = PCI_SUBSYSTEM_ID_VESA, .bar[1] = VESA_MEM_ADDR | PCI_BASE_ADDRESS_SPACE_MEMORY, + .bar_size[1] = VESA_MEM_SIZE, }; static struct framebuffer vesafb; @@ -56,7 +57,6 @@ struct framebuffer *vesa__init(struct kvm *kvm) vesa_pci_device.irq_line = line; vesa_base_addr = ioport__register(IOPORT_EMPTY, &vesa_io_ops, IOPORT_SIZE, NULL); vesa_pci_device.bar[0] = vesa_base_addr | PCI_BASE_ADDRESS_SPACE_IO; - vesa_pci_device.bar_size[0] = VESA_MEM_SIZE; pci__register(&vesa_pci_device, dev); mem = mmap(NULL, VESA_MEM_SIZE, PROT_RW, MAP_ANON_NORESERVE, -1, 0); diff --git a/tools/kvm/pci.c b/tools/kvm/pci.c index fd19b73..0449aca 100644 --- a/tools/kvm/pci.c +++ b/tools/kvm/pci.c @@ -95,7 +95,7 @@ static bool pci_config_data_out(struct ioport *ioport, struct kvm *kvm, u16 port offset = start + (pci_config_address.register_number << 2); if (offset < sizeof(struct pci_device_header)) { void *p = pci_devices[dev_num]; - u8 bar = offset - PCI_BAR_OFFSET(0); + u8 bar = offset - PCI_BAR_OFFSET(0) / (sizeof(u32)); u32 sz = PCI_IO_SIZE; if (bar < 6 && pci_devices[dev_num]->bar_size[bar])
This patch makes BAR 1 16k, instead of BAR0 - which is the PIO bar. Signed-off-by: Sasha Levin <levinsasha928@gmail.com> --- tools/kvm/hw/vesa.c | 2 +- tools/kvm/pci.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)