[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [BUG] After upgrade to Xen 4.12.0 iommu=no-igfx
Sorry -- got a bit distracted yesterday. Attached is the log with only your latest patch attached. Interestingly enough the box booted fine without screen artifacts. So I guess we're getting closer... Thanks for all the help! Thanks, Roman. On Tue, Jul 30, 2019 at 2:21 AM Roger Pau Monné <roger.pau@xxxxxxxxxx> wrote: > > Ping? > > It would be very helpful for us in order to get this sorted out if you > could answer the questions below and try the new debug patch :). > > On Fri, Jul 26, 2019 at 11:35:45AM +0200, Roger Pau Monné wrote: > > On Thu, Jul 25, 2019 at 05:47:19PM -0700, Roman Shaposhnik wrote: > > > Hi Roger! > > > > > > With your patch (and build as a debug build) Xen crashes on boot > > > (which I guess was the point of your BUG_ON statement). > > > > Yes, that's very weird, seems like entries are not added to the iommu > > page tables but I have no idea why, AFAICT this works fine on my > > system. > > > > Do you have any patches on top of RELEASE-4.12.0? > > > > I have another patch with more verbose output, could you give it a > > try? It's maybe going to be more chatty than the previous one. > > > > I'm sorry to keep you testing stuff, but since I cannot reproduce this > > locally I have to rely on you to provide the debug output. > > > > Thanks, Roger. > > ---8<--- > > diff --git a/xen/arch/x86/mm/p2m.c b/xen/arch/x86/mm/p2m.c > > index b9bbb8f485..75f8359a99 100644 > > --- a/xen/arch/x86/mm/p2m.c > > +++ b/xen/arch/x86/mm/p2m.c > > @@ -1331,7 +1331,7 @@ int set_identity_p2m_entry(struct domain *d, unsigned > > long gfn_l, > > > > if ( !paging_mode_translate(p2m->domain) ) > > { > > - if ( !need_iommu_pt_sync(d) ) > > + if ( !has_iommu_pt(d) ) > > return 0; > > return iommu_legacy_map(d, _dfn(gfn_l), _mfn(gfn_l), PAGE_ORDER_4K, > > IOMMUF_readable | IOMMUF_writable); > > @@ -1422,7 +1422,7 @@ int clear_identity_p2m_entry(struct domain *d, > > unsigned long gfn_l) > > > > if ( !paging_mode_translate(d) ) > > { > > - if ( !need_iommu_pt_sync(d) ) > > + if ( !has_iommu_pt(d) ) > > return 0; > > return iommu_legacy_unmap(d, _dfn(gfn_l), PAGE_ORDER_4K); > > } > > diff --git a/xen/drivers/passthrough/iommu.c > > b/xen/drivers/passthrough/iommu.c > > index 117b869b0c..214c5d515f 100644 > > --- a/xen/drivers/passthrough/iommu.c > > +++ b/xen/drivers/passthrough/iommu.c > > @@ -291,8 +291,18 @@ int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn, > > unsigned long i; > > int rc = 0; > > > > +if (dfn_x(dfn) >= 0x8d800 && dfn_x(dfn) < 0x90000 ) > > +{ > > + printk("<RMRR> iommu_map %#lx\n", dfn_x(dfn)); > > + process_pending_softirqs(); > > +} > > + > > if ( !iommu_enabled || !hd->platform_ops ) > > +{ > > + printk("iommu_enabled: %d platform_ops %p\n", > > + iommu_enabled, hd->platform_ops); > > return 0; > > +} > > > > ASSERT(IS_ALIGNED(dfn_x(dfn), (1ul << page_order))); > > ASSERT(IS_ALIGNED(mfn_x(mfn), (1ul << page_order))); > > diff --git a/xen/drivers/passthrough/vtd/iommu.c > > b/xen/drivers/passthrough/vtd/iommu.c > > index 50a0e25224..8c3fcb50ae 100644 > > --- a/xen/drivers/passthrough/vtd/iommu.c > > +++ b/xen/drivers/passthrough/vtd/iommu.c > > @@ -2009,12 +2009,33 @@ static int rmrr_identity_mapping(struct domain *d, > > bool_t map, > > if ( !map ) > > return -ENOENT; > > > > +printk("<RMRR> mapping %#lx - %#lx\n", base_pfn, end_pfn); > > while ( base_pfn < end_pfn ) > > { > > int err = set_identity_p2m_entry(d, base_pfn, p2m_access_rw, flag); > > + mfn_t mfn; > > + unsigned int f; > > > > if ( err ) > > return err; > > + > > +err = intel_iommu_lookup_page(d, _dfn(base_pfn), &mfn, &f); > > +if ( err ) > > +{ > > + printk("intel_iommu_lookup_page err: %d\n", err); > > + BUG(); > > +} > > +if ( base_pfn != mfn_x(mfn) ) > > +{ > > + printk("base_pfn: %#lx mfn: %#lx\n", base_pfn, mfn_x(mfn)); > > + BUG(); > > +} > > +if ( f != (IOMMUF_readable | IOMMUF_writable) ) > > +{ > > + printk("flags: %#x\n", f); > > + BUG(); > > +} > > + > > base_pfn++; > > } > > > > @@ -2263,6 +2284,7 @@ static void __hwdom_init setup_hwdom_rmrr(struct > > domain *d) > > u16 bdf; > > int ret, i; > > > > +printk("<RMRR> setting up regions\n"); > > pcidevs_lock(); > > for_each_rmrr_device ( rmrr, bdf, i ) > > { > > > > _______________________________________________ > > Xen-devel mailing list > > Xen-devel@xxxxxxxxxxxxxxxxxxxx > > https://lists.xenproject.org/mailman/listinfo/xen-devel Attachment:
xen.txt _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxxx https://lists.xenproject.org/mailman/listinfo/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |