Re: [RFC PATCH-for-5.2] hw/s390x/pci: Fix endianness issue

[Cornelia Huck]

On Tue, 17 Nov 2020 13:01:15 +0100
Philippe Mathieu-Daudé <philmd@redhat.com> wrote:

> Fix an endianness issue reported by Cornelia:
> 
> > s390x tcg guest on x86, virtio-pci devices are not detected. The
> > relevant feature bits are visible to the guest. Same breakage with
> > different guest kernels.
> > KVM guests and s390x tcg guests on s390x are fine.  
> 
> Fixes: 28dc86a0729 ("s390x/pci: use a PCI Group structure")
> Reported-by: Cornelia Huck <cohuck@redhat.com>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
> RFC because review-only patch, untested
> ---
>  hw/s390x/s390-pci-inst.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/hw/s390x/s390-pci-inst.c b/hw/s390x/s390-pci-inst.c
> index 58cd041d17f..cfb54b4d8ec 100644
> --- a/hw/s390x/s390-pci-inst.c
> +++ b/hw/s390x/s390-pci-inst.c
> @@ -305,7 +305,7 @@ int clp_service_call(S390CPU *cpu, uint8_t r2, uintptr_t 
> ra)
>          ClpReqQueryPciGrp *reqgrp = (ClpReqQueryPciGrp *)reqh;
>          S390PCIGroup *group;
>  
> -        group = s390_group_find(reqgrp->g);
> +        group = s390_group_find(ldl_p(&reqgrp->g));
>          if (!group) {
>              /* We do not allow access to unknown groups */
>              /* The group must have been obtained with a vfio device */

Hmm... when I tested this change on top of 28dc86a0729, it worked; when
tested on top of my s390-fixes branch, I get

operand exception: 0015 ilc:3 [#1] SMP 
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.8.0-rc3+ #174
Hardware name: QEMU 2964 QEMU (KVM/Linux)
Krnl PSW : 0704d00180000000 00000000001727be (zpci_mod_fc+0x2e/0xf8)
           R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:1 PM:0 RI:0 EA:3
Krnl GPRS: 000003d080fb8040 8002000000000004 8002000000000004 000003e00000bb30
           000003e00000bb2f 0000000040ffffff 000000003efa8804 0000000000000002
           000003e00000bb2f 8002000000000004 0000000000000000 000000003f1b4400
           000000003f41c000 0000000000000000 000003e00000ba88 000003e00000ba20
Krnl Code: 00000000001727ae: e3e0f0980024       stg     %r14,152(%r15)
           00000000001727b4: b9040084           lgr     %r8,%r4
          #00000000001727b8: e310300000d0       mpcifc  %r1,0(%r3)
          >00000000001727be: b22200b0           ipm     %r11
           00000000001727c2: 88b0001c           srl     %r11,28
           00000000001727c6: 42b0f0a7           stc     %r11,167(%r15)
           00000000001727ca: eb110018000c       srlg    %r1,%r1,24
           00000000001727d0: 42104000           stc     %r1,0(%r4)
Call Trace:
 [<00000000001727be>] zpci_mod_fc+0x2e/0xf8 
 [<000000000016d364>] zpci_register_ioat+0x64/0x88 
 [<0000000000170532>] zpci_dma_init_device+0x13a/0x1b8 
 [<000000000016e008>] zpci_enable_device+0x48/0x70 
 [<000000000016e1b2>] zpci_create_device+0x14a/0x170 
 [<0000000000171224>] clp_add_pci_device+0x504/0x528 
 [<0000000000170c46>] clp_list_pci+0xf6/0x1d0 
 [<00000000001712f4>] clp_scan_pci_devices+0x44/0x68 
 [<00000000011e859e>] pci_base_init+0x12e/0x1a0 
 [<00000000011ddf92>] do_one_initcall+0x132/0x2e0 
 [<00000000011de3cc>] kernel_init_freeable+0x28c/0x308 
 [<0000000000cfc822>] kernel_init+0x22/0x150 
 [<0000000000d08f60>] ret_from_fork+0x24/0x28 
INFO: lockdep is turned off.
Last Breaking-Event-Address:
 [<000000000016d35e>] zpci_register_ioat+0x5e/0x88

(Yeah, I know it's not the newest guest kernel.)

This patch probably uncovers further endianness problems in the
following zpci patches. Looking.