Re: [qemu-s390x] [Qemu-devel] [PATCH 1/3] s390x/pci: fixup the code walking IOMMU tables

[Yi Min Zhao]

在 2018/1/31 下午3:42, Thomas Huth 写道:
> On 30.01.2018 10:47, Yi Min Zhao wrote:
> > Current s390x PCI IOMMU code is lack of flags' checking, including:
> > 1) protection bit
> > 2) table length
> > 3) table offset
> > 4) intermediate tables' invalid bit
> > 5) format control bit
> >
> > This patch introduces a new struct named S390IOTLBEntry, and makes up
> > these missed checkings. At the same time, inform the guest with the
> > corresponding error number when the check fails.
> >
> > Reviewed-by: Pierre Morel <address@hidden>
> > Signed-off-by: Yi Min Zhao <address@hidden>
> > ---
> >   hw/s390x/s390-pci-bus.c  | 223 ++++++++++++++++++++++++++++++++++++++---------
> >   hw/s390x/s390-pci-bus.h  |  10 +++
> >   hw/s390x/s390-pci-inst.c |  10 ---
> >   3 files changed, 190 insertions(+), 53 deletions(-)
> [...]
> > @@ -374,26 +511,26 @@ static IOMMUTLBEntry 
> > s390_translate_iommu(IOMMUMemoryRegion *mr, hwaddr addr,
> >       DPRINTF("iommu trans addr 0x%" PRIx64 "\n", addr);
> >   
> >       if (addr < iommu->pba || addr > iommu->pal) {
> > -        return ret;
> > +        error = ERR_EVENT_OORANGE;
> > +        goto err;
> >       }
> >   
> > -    pte = s390_guest_io_table_walk(s390_pci_get_table_origin(iommu->g_iota),
> > -                                   addr);
> > -    if (!pte) {
> > -        return ret;
> > -    }
> > +    error = s390_guest_io_table_walk(iommu->g_iota, addr, &entry);
> >   
> > -    flags = pte & ZPCI_PTE_FLAG_MASK;
> > -    ret.iova = addr;
> > -    ret.translated_addr = pte & ZPCI_PTE_ADDR_MASK;
> > -    ret.addr_mask = 0xfff;
> > +    ret.iova = entry.iova;
> > +    ret.translated_addr = entry.translated_addr;
> > +    ret.addr_mask = entry.len - 1;
> > +    ret.perm = entry.perm;
> >   
> > -    if (flags & ZPCI_PTE_INVALID) {
> > -        ret.perm = IOMMU_NONE;
> > -    } else {
> > -        ret.perm = IOMMU_RW;
> > +    if ((flag != IOMMU_NONE) && !(flag & ret.perm)) {
> You could drop the parentheses around "flag != IOMMU_NONE".
OK. Will update.
> For the rest of the patch: Sorry, can't review due to missing PCI spec :-(
Thanks for your review anyway!
>   Thomas