[Qemu-devel] Re: [PATCH 14/15] megasas: LSI Megaraid SAS emulation

[Stefan Hajnoczi]

On Thu, Nov 25, 2010 at 2:50 PM, Hannes Reinecke <address@hidden> wrote:
> On 11/25/2010 03:36 PM, Stefan Hajnoczi wrote:
>> On Wed, Nov 24, 2010 at 11:16 AM, Hannes Reinecke <address@hidden> wrote:
>>> +static int megasas_pd_get_info_submit(SCSIDevice * sdev, int lun,
>>> +                                      struct megasas_cmd_t *cmd)
>>> +{
>>> +    struct mfi_pd_info * info = cmd->iov_buf;
>>> +    uint8_t cmdbuf[6];
>>> +    SCSIRequest *req;
>>> +
>>> +    if (info->inquiry_data[4] == 0) {
>>> +        /* Additional length is zero, resubmit */
>>> +        megasas_setup_inquiry(cmdbuf, 0, info->inquiry_data,
>>> +                              sizeof(info->inquiry_data));
>>> +        req = sdev->info->alloc_req(sdev, (uint32_t) -1, lun);
>>> +        if (!req) {
>>> +            return MFI_STAT_FLASH_ALLOC_FAIL;
>>> +        }
>>> +        DPRINTF_DCMD("PD get info submit std inquiry to dev %d\n", lun);
>>> +        req->hba_private = cmd;
>>> +        if (cmd->sdev->info->send_command(req, cmdbuf) > 0)
>>> +            cmd->sdev->info->read_data(req);
>>> +        return MFI_STAT_INVALID_STATUS;
>>
>> Here...
>>
>>> +    } else if (info->vpd_page83[3] == 0) {
>>> +        /* Additional length is zero, resubmit */
>>> +        megasas_setup_inquiry(cmdbuf, 0x83,(uint8_t *)info->vpd_page83,
>>> +                              sizeof(info->vpd_page83));
>>> +        req = sdev->info->alloc_req(sdev, (uint32_t) -1, lun);
>>> +        if (!req) {
>>> +            return MFI_STAT_FLASH_ALLOC_FAIL;
>>> +        }
>>> +        DPRINTF_DCMD("PD get info submit vpd inquiry to dev %d\n", lun);
>>> +        req->hba_private = cmd;
>>> +        if (cmd->sdev->info->send_command(req, cmdbuf) > 0)
>>> +            cmd->sdev->info->read_data(req);
>>> +        return MFI_STAT_INVALID_STATUS;
>>
>> ...and here I can't tell for sure if an error status is returned
>> intentionally or whether this is an if statement without {}-related
>> bug.
>>
>> On one hand it looks like it could be intentional.  On the other hand
>> we went through the trouble of sending an inquiry command that may
>> have succeeded but we're still returning an error status.
>>
>
> MFI_STAT_INVALID_STATUS is _not_ an error status. It's used
> internally to signal that a command is still being processed.
> ->read_data() itself doesn't have a meaningful return code;
> we need to wait for the ->complete callback to get results.
> So we're returning MFI_STAT_INVALID_STATUS to signal this condition.
>
> So that's ok.

Okay, thanks for explaining.

Stefan