Re: [RFC PATCH 1/1] Add pci_hole_min_size

[Don Slutz <dslutz@verizon.com>]

On 03/03/14 11:07, Boris Ostrovsky wrote:
> On 03/03/2014 10:30 AM, Don Slutz wrote:
>> On 02/28/14 17:07, Boris Ostrovsky wrote:
>>>
>>>> diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
>>>> index a604cd8..24ceac6 100644
>>>> --- a/tools/libxl/libxl_create.c
>>>> +++ b/tools/libxl/libxl_create.c
>>>> @@ -388,13 +388,15 @@ int libxl__domain_build(libxl__gc *gc,
>>>>           vments[4] = "start_time";
>>>>           vments[5] = libxl__sprintf(gc, "%lu.%02d", start_time.tv_sec,(int)start_time.tv_usec/10000);
>>>>   -        localents = libxl__calloc(gc, 7, sizeof(char *));
>>>> +        localents = libxl__calloc(gc, 9, sizeof(char *));
>>>>           localents[0] = "platform/acpi";
>>>>           localents[1] = libxl_defbool_val(info->u.hvm.acpi) ? "1" : "0";
>>>>           localents[2] = "platform/acpi_s3";
>>>>           localents[3] = libxl_defbool_val(info->u.hvm.acpi_s3) ? "1" : "0";
>>>>           localents[4] = "platform/acpi_s4";
>>>>           localents[5] = libxl_defbool_val(info->u.hvm.acpi_s4) ? "1" : "0";
>>>> +        localents[6] = "platform/pci_hole_min_size";
>>>> +        localents[7] = libxl__sprintf(gc, "%llu", (unsigned long long)info->u.hvm.pci_hole_min_size);
>>>
>>> Do you want to always store this parameter? There is a default already (HVM_BELOW_4G_MMIO_LENGTH) so if it's not set in the config file it may be safe to omit it.
>>>
>>
>> I do not always need to store it.  Since none of the rest of these are conditional stores, I just followed them.  Since this is the minimum size, I can add a check on pci_hole_min_size > HVM_BELOW_4G_MMIO_LENGTH here and skip the xenstore write if you want.
>
> If you decide to do this I think the better place may be in libxl__build_post().
>

I lean to not doing it.

>>
>>> ...
>>>
>>>
>>>> diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c
>>>> index 4fc46eb..fe247ee 100644
>>>> --- a/tools/libxl/xl_cmdimpl.c
>>>> +++ b/tools/libxl/xl_cmdimpl.c
>>>> @@ -1025,6 +1025,12 @@ static void parse_config_data(const char *config_source,
>>>>           xlu_cfg_get_defbool(config, "hpet", &b_info->u.hvm.hpet, 0);
>>>>           xlu_cfg_get_defbool(config, "vpt_align", &b_info->u.hvm.vpt_align, 0);
>>>>   +        if (!xlu_cfg_get_long(config, "pci_hole_min_size", &l, 0)) {
>>>> +            b_info->u.hvm.pci_hole_min_size = (uint64_t) l;
>>>> +            if (dom_info->debug)
>>>> +                fprintf(stderr, "pci_hole_min_size: %llu\n", (unsigned long long) b_info->u.hvm.pci_hole_min_size);
>>>> +        }
>>>
>>> You probably want to set b_info->u.hvm.pci_hole_min_size to 0 (or HVM_BELOW_4G_MMIO_LENGTH?) in case it's not specified in libxl__domain_build_info_setdefault().
>>
>> Since 0 is a valid value, I do not think that libxl__domain_build_info_setdefault() needs to do any thing. Should I be
>> specifying a default in the idl of 0?
>
>
> What I meant (but apparently not what I wrote) was that if config file doesn't have "pci_hole_min_size" then b_info->u.hvm.pci_hole_min_size will be left uninitialized and I don't know whether we can assume that it will be zero.
>

As expected, changing the .idl to have an init:

- ("pci_hole_min_size",uint64),
+ ("pci_hole_min_size",UInt(64, init_val = 1)),

The 1 is setup in the generated code routine libxl_domain_build_info_init_type().  I plan to change this to 0.

Since libxl__domain_build_info_setdefault() is called after parse_config_data() it is not simple to code.

If you do not like the .idl to have a 0, (i.e. like LIBXL_MEMKB_DEFAULT) then I will need to add a LIBXL_PCI_HOLE_MIN_DEFAULT define (like #define LIBXL_PCI_HOLE_MIN_DEFAULT ~0ULL). And then use it.

    -Don Slutz

> -boris
>
>
>>
>>    -Don Slutz
>>>
>>> -boris
>>>
>>>
>>>> +
>>>>           if (!xlu_cfg_get_long(config, "timer_mode", &l, 1)) {
>>>>               const char *s = libxl_timer_mode_to_string(l);
>>>>               fprintf(stderr, "WARNING: specifying \"timer_mode\" as an integer is deprecated. "
>>>
>>
>