Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-17 Thread Konrad Rzeszutek Wilk
On Mon, Sep 17, 2012 at 03:52:20PM +0100, Stefano Stabellini wrote: > On Mon, 17 Sep 2012, Konrad Rzeszutek Wilk wrote: > > > > start_dma_addr = xen_virt_to_bus(xen_io_tlb_start); > > > > - if (!after_bootmem) > > > > + rc = 0; > > > ^ > > > why does this change belong to t

Re: Is: [PATCH 11/10] xen/swiotlb: For early initialization, return zero on success. Was: Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-17 Thread Stefano Stabellini
On Mon, 17 Sep 2012, Konrad Rzeszutek Wilk wrote: > On Mon, Sep 17, 2012 at 10:23:15AM -0400, Konrad Rzeszutek Wilk wrote: > > > > start_dma_addr = xen_virt_to_bus(xen_io_tlb_start); > > > > - if (!after_bootmem) > > > > + rc = 0; > > > ^ > > > why does this change belong t

Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-17 Thread Stefano Stabellini
On Mon, 17 Sep 2012, Konrad Rzeszutek Wilk wrote: > > > start_dma_addr = xen_virt_to_bus(xen_io_tlb_start); > > > - if (!after_bootmem) > > > + rc = 0; > > ^ > > why does this change belong to this patch? > > > > > > I took that out of the this patch, so it is now: > > > >From c5bc5502a

Is: [PATCH 11/10] xen/swiotlb: For early initialization, return zero on success. Was: Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-17 Thread Konrad Rzeszutek Wilk
On Mon, Sep 17, 2012 at 10:23:15AM -0400, Konrad Rzeszutek Wilk wrote: > > > start_dma_addr = xen_virt_to_bus(xen_io_tlb_start); > > > - if (!after_bootmem) > > > + rc = 0; > > ^ > > why does this change belong to this patch? > > > > > > I took that out of the this patch, so it is now: >

Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-17 Thread Konrad Rzeszutek Wilk
> > start_dma_addr = xen_virt_to_bus(xen_io_tlb_start); > > - if (!after_bootmem) > > + rc = 0; > ^ > why does this change belong to this patch? > > I took that out of the this patch, so it is now: >From c5bc5502abc0f70b682c0f2a70d08e6319825163 Mon Sep 17 00:00:00 2001 From: Konra

Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-14 Thread Konrad Rzeszutek Wilk
On Fri, Sep 14, 2012 at 05:10:48PM +0100, Stefano Stabellini wrote: > On Mon, 10 Sep 2012, Konrad Rzeszutek Wilk wrote: > > When PCI IOMMUs are initialized it is after after_bootmem but > > before a lot of "other" subsystems are initialized. As such > > the check for after_bootmem is incorrect and

Re: [PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-14 Thread Stefano Stabellini
On Mon, 10 Sep 2012, Konrad Rzeszutek Wilk wrote: > When PCI IOMMUs are initialized it is after after_bootmem but > before a lot of "other" subsystems are initialized. As such > the check for after_bootmem is incorrect and we should > just use a parameter to define whether we are early or late. >

[PATCH 10/10] xen/swiotlb: Depending on after_bootmem is not correct.

2012-09-10 Thread Konrad Rzeszutek Wilk
When PCI IOMMUs are initialized it is after after_bootmem but before a lot of "other" subsystems are initialized. As such the check for after_bootmem is incorrect and we should just use a parameter to define whether we are early or late. This solves this bootup problem: __ex_table already sorted,