qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Alistair Francis <alistair.francis@xilinx.com>
To: Peter Crosthwaite <crosthwaitepeter@gmail.com>
Cc: "Edgar Iglesias" <edgar.iglesias@xilinx.com>,
	"Peter Maydell" <peter.maydell@linaro.org>,
	"qemu-devel@nongnu.org Developers" <qemu-devel@nongnu.org>,
	"Alistair Francis" <alistair.francis@xilinx.com>,
	"Sai Pavan Boddu" <saipava@xilinx.com>,
	"John Snow" <jsnow@redhat.com>,
	"Andreas Färber" <afaerber@suse.de>
Subject: Re: [Qemu-devel] [PATCH v3 2/4] ahci.c: Don't assume AHCIState's parent is AHCIPCIState
Date: Thu, 27 Aug 2015 15:58:50 -0700	[thread overview]
Message-ID: <CAKmqyKNMDBbiNgKUe6f0SUVCmLPodZiST032LQFn4PSX6xmFEQ@mail.gmail.com> (raw)
In-Reply-To: <CAPokK=oca9-XFLKaudTyupACFaqkMhCpn6-AVJESU_sSKbaKCA@mail.gmail.com>

()

On Thu, Aug 27, 2015 at 3:36 PM, Peter Crosthwaite
<crosthwaitepeter@gmail.com> wrote:
> On Thu, Aug 27, 2015 at 3:06 PM, Alistair Francis
> <alistair.francis@xilinx.com> wrote:
>> On Thu, Aug 27, 2015 at 2:28 PM, Peter Crosthwaite
>> <crosthwaitepeter@gmail.com> wrote:
>>> On Thu, Aug 27, 2015 at 1:47 PM, Alistair Francis
>>> <alistair.francis@xilinx.com> wrote:
>>>> The AHCIState strucut can either have AHCIPCIState or SysbusAHCIState
>>>
>>> "struct"
>>
>> Good catch, will fix
>>
>>>
>>>> as a parent. The ahci_irq_lower() and ahci_irq_raise() functions
>>>> assume that it is always AHCIPCIState, which is not always the
>>>> case, which causes a seg fault. Verify what the container of AHCIState
>>>> is before setting the PCIDevice struct.
>>>>
>>>> Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
>>>> ---
>>>>
>>>>  hw/ide/ahci.c |   29 +++++++++++++++++++++++------
>>>>  hw/ide/ahci.h |    2 ++
>>>>  2 files changed, 25 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/hw/ide/ahci.c b/hw/ide/ahci.c
>>>> index 02d85fa..b45b21d 100644
>>>> --- a/hw/ide/ahci.c
>>>> +++ b/hw/ide/ahci.c
>>>> @@ -121,9 +121,17 @@ static uint32_t  ahci_port_read(AHCIState *s, int port, int offset)
>>>>
>>>>  static void ahci_irq_raise(AHCIState *s, AHCIDevice *dev)
>>>>  {
>>>> -    AHCIPCIState *d = container_of(s, AHCIPCIState, ahci);
>>>> -    PCIDevice *pci_dev =
>>>> -        (PCIDevice *)object_dynamic_cast(OBJECT(d), TYPE_PCI_DEVICE);
>>>
>>> Why can't you just convert this to use use s->container? ...
>>>
>>>> +    DeviceState *dev_state = DEVICE(s->container);
>>>
>>> Unneeded cast. s->container is already DeviceState, but I'm not sure
>>> you need this local variable at all.
>>
>> I'll remove the cast. The variable is just included to save line space
>> (and complexity).
>>
>>>
>>>> +    PCIDevice *pci_dev = NULL;
>>>> +    ObjectClass *ret;
>>>> +
>>>> +    /* Check is AHCIState's parent is SysbusAHCIState or AHCIPCIState */
>>>> +    ret = object_class_dynamic_cast(object_get_class(OBJECT(dev_state)),
>>>> +                                    TYPE_PCI_DEVICE);
>>>
>>> ... I don't understand why this needs to be done as a class cast, does
>>> the original object cast approach not work?
>>
>> The problem with the standard ones is that they contain asserts. So you can't
>> do a test to see if the cast is possible, which is what I want to do.
>>
>
> But the original code uses object_dynamic_cast which should be a
> non-asserting cast attempt. What was the assertion you were getting?

Oh! I see what you mean. I can use the object_dynamic_cast() instead.
I thought you meant to use the macros that are defined.

If there are no other comments I will re-send the series.

Thanks,

Alistair

>
> Regards,
> Peter
>
>> Thanks,
>>
>> Alistair
>>
>>>
>>> Regards,
>>> Peter
>>>
>>>> +    if (ret) {
>>>> +        /* AHCIState parent is AHCIPCIState */
>>>> +        pci_dev = PCI_DEVICE(dev_state);
>>>> +    }
>>>>
>>>>      DPRINTF(0, "raise irq\n");
>>>>
>>>> @@ -136,9 +144,17 @@ static void ahci_irq_raise(AHCIState *s, AHCIDevice *dev)
>>>>
>>>>  static void ahci_irq_lower(AHCIState *s, AHCIDevice *dev)
>>>>  {
>>>> -    AHCIPCIState *d = container_of(s, AHCIPCIState, ahci);
>>>> -    PCIDevice *pci_dev =
>>>> -        (PCIDevice *)object_dynamic_cast(OBJECT(d), TYPE_PCI_DEVICE);
>>>> +    DeviceState *dev_state = DEVICE(s->container);
>>>> +    PCIDevice *pci_dev = NULL;
>>>> +    ObjectClass *ret;
>>>> +
>>>> +    /* Check is AHCIState's parent is SysbusAHCIState or AHCIPCIState */
>>>> +    ret = object_class_dynamic_cast(object_get_class(OBJECT(dev_state)),
>>>> +                                    TYPE_PCI_DEVICE);
>>>> +    if (ret) {
>>>> +        /* AHCIState parent is AHCIPCIState */
>>>> +        pci_dev = PCI_DEVICE(dev_state);
>>>> +    }
>>>>
>>>>      DPRINTF(0, "lower irq\n");
>>>>
>>>> @@ -1436,6 +1452,7 @@ void ahci_init(AHCIState *s, DeviceState *qdev, AddressSpace *as, int ports)
>>>>      s->as = as;
>>>>      s->ports = ports;
>>>>      s->dev = g_new0(AHCIDevice, ports);
>>>> +    s->container = qdev;
>>>>      ahci_reg_init(s);
>>>>      /* XXX BAR size should be 1k, but that breaks, so bump it to 4k for now */
>>>>      memory_region_init_io(&s->mem, OBJECT(qdev), &ahci_mem_ops, s,
>>>> diff --git a/hw/ide/ahci.h b/hw/ide/ahci.h
>>>> index c055d6b..c9b3805 100644
>>>> --- a/hw/ide/ahci.h
>>>> +++ b/hw/ide/ahci.h
>>>> @@ -287,6 +287,8 @@ struct AHCIDevice {
>>>>  };
>>>>
>>>>  typedef struct AHCIState {
>>>> +    DeviceState *container;
>>>> +
>>>>      AHCIDevice *dev;
>>>>      AHCIControlRegs control_regs;
>>>>      MemoryRegion mem;
>>>> --
>>>> 1.7.1
>>>>
>>>
>

  reply	other threads:[~2015-08-27 22:59 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-08-27 20:47 [Qemu-devel] [PATCH v3 0/4] xlnx-zynqmp: Connect the AHCI SATA device Alistair Francis
2015-08-27 20:47 ` [Qemu-devel] [PATCH v3 1/4] ahci: Separate the AHCI state structure into the header Alistair Francis
2015-08-27 20:47 ` [Qemu-devel] [PATCH v3 2/4] ahci.c: Don't assume AHCIState's parent is AHCIPCIState Alistair Francis
2015-08-27 21:28   ` Peter Crosthwaite
2015-08-27 22:06     ` Alistair Francis
2015-08-27 22:36       ` Peter Crosthwaite
2015-08-27 22:58         ` Alistair Francis [this message]
2015-08-27 23:25           ` Peter Crosthwaite
2015-08-27 20:47 ` [Qemu-devel] [PATCH v3 3/4] xlnx-zynqmp.c: Convert some of the error_propagate() calls to error_abort Alistair Francis
2015-08-27 21:32   ` Peter Crosthwaite
2015-08-27 21:50     ` Alistair Francis

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAKmqyKNMDBbiNgKUe6f0SUVCmLPodZiST032LQFn4PSX6xmFEQ@mail.gmail.com \
    --to=alistair.francis@xilinx.com \
    --cc=afaerber@suse.de \
    --cc=crosthwaitepeter@gmail.com \
    --cc=edgar.iglesias@xilinx.com \
    --cc=jsnow@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    --cc=saipava@xilinx.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).