From: Boris Ostrovsky <boris.ostrovsky@oracle.com>
To: Jan Beulich <JBeulich@suse.com>
Cc: wei.liu2@citrix.com, andrew.cooper3@citrix.com,
ian.jackson@eu.citrix.com, xen-devel@lists.xen.org,
julien.grall@arm.com, zhaoshenglong@huawei.com,
roger.pau@citrix.com
Subject: Re: [PATCH v1 02/20] acpi/hvmloader: Move acpi_info initialization out of ACPI code
Date: Mon, 1 Aug 2016 10:06:19 -0400 [thread overview]
Message-ID: <40889253-ce31-eebe-cf78-8ad27dd31f82@oracle.com> (raw)
In-Reply-To: <579F3C0A02000078001012E7@prv-mh.provo.novell.com>
On 08/01/2016 06:09 AM, Jan Beulich wrote:
>>>> On 08.07.16 at 18:14, <boris.ostrovsky@oracle.com> wrote:
>> On 07/08/2016 11:11 AM, Jan Beulich wrote:
>>>>>> On 08.07.16 at 16:39, <boris.ostrovsky@oracle.com> wrote:
>>>> On 07/08/2016 06:10 AM, Jan Beulich wrote:
>>>>>> @@ -615,20 +593,10 @@ void acpi_build_tables(struct acpi_config *config,
>>>> unsigned int physical)
>>>>>> offsetof(struct acpi_20_rsdp, extended_checksum),
>>>>>> sizeof(struct acpi_20_rsdp));
>>>>>>
>>>>>> - if ( !new_vm_gid(acpi_info) )
>>>>>> + if ( !new_vm_gid(&config->ainfo) )
>>>>>> goto oom;
>>>>>>
>>>>>> - acpi_info->com1_present = uart_exists(0x3f8);
>>>>>> - acpi_info->com2_present = uart_exists(0x2f8);
>>>>>> - acpi_info->lpt1_present = lpt_exists(0x378);
>>>>>> - acpi_info->hpet_present = hpet_exists(ACPI_HPET_ADDRESS);
>>>>>> - acpi_info->pci_min = pci_mem_start;
>>>>>> - acpi_info->pci_len = pci_mem_end - pci_mem_start;
>>>>>> - if ( pci_hi_mem_end > pci_hi_mem_start )
>>>>>> - {
>>>>>> - acpi_info->pci_hi_min = pci_hi_mem_start;
>>>>>> - acpi_info->pci_hi_len = pci_hi_mem_end - pci_hi_mem_start;
>>>>>> - }
>>>>>> + *(struct acpi_info *)config->ainfop = config->ainfo;
>>>>> With your new separation of responsibilities - does this really
>>>>> belong here rather than in the caller?
>>>> I think it should be done here: when the call returns all tables are
>>>> already in memory. If the caller wants to load those tables separately
>>>> (as probably the toolstack will) then it can simply copy it as a blob.
>>> But this structure isn't part of the ACPI tables, and by not doing
>>> it here (a) at least some of the intended callers may be able to
>>> get away without the ugly cast and (b) the field now named
>>> ainfop wouldn't be needed either afaict.
>>
>> I probably didn't use right terminology. This is not a table, but an AML
>> piece?
> Clearly not. This is data structure we define ourselves, which only
> gets used by AML code.
>
>> In any case, it's something that is ACPI-specific and I was
>> hoping we wouldn't need to expose this to the caller.
> That would imo be a relevant argument only if the structure
> type was indeed private to a single (sub-)component.
As I said below, I only expose this structure to callers for convenience.
How about I keep acpi_info private to the builder and instead define new
structure that will pass necessary information to the builder so that it
can fill acpi_info internally and copy it to memory?
This will also have side effect of allowing us to keep acpi_info and
tables in contiguous memory. We do this now implicitly for hvmloader
(RESERVED_MEMORY_DYNAMIC_START = ACPI_INFO_PHYSICAL_ADDRESS + PAGE_SIZE)
so nothing should change from that side but from tools perspective it's
easier to deal with a single ACPI blob.
-boris
>
>> The fact that it
>> is passed in the right format in struct acpi_info is a happy
>> coincidence. We may change it in the future (and so perhaps I should
>> drop the comment in libacpi.h about "This must match the
>> Field("BIOS"....) definition in the DSDT.")
> Definitely not: The two absolutely have to remain in sync. They're
> C and AML representations of the same thing.
>
> Jan
>
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
next prev parent reply other threads:[~2016-08-01 14:06 UTC|newest]
Thread overview: 104+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-07-05 19:04 [PATCH v1 00/20] Make ACPI builder available to components other than hvmloader Boris Ostrovsky
2016-07-05 19:05 ` [PATCH v1 01/20] hvmloader: Provide hvmloader_acpi_build_tables() Boris Ostrovsky
2016-07-06 14:47 ` Konrad Rzeszutek Wilk
2016-07-08 9:52 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 02/20] acpi/hvmloader: Move acpi_info initialization out of ACPI code Boris Ostrovsky
2016-07-07 16:58 ` Ian Jackson
2016-07-07 17:09 ` Boris Ostrovsky
2016-07-07 17:15 ` Wei Liu
2016-07-07 17:45 ` Boris Ostrovsky
2016-07-08 15:06 ` Konrad Rzeszutek Wilk
2016-07-08 15:50 ` Ian Jackson
2016-07-08 15:57 ` Boris Ostrovsky
2016-07-08 16:21 ` Ian Jackson
2016-07-11 12:10 ` Wei Liu
2016-07-11 14:47 ` Lars Kurth
2016-07-11 14:54 ` Konrad Rzeszutek Wilk
2016-07-11 15:06 ` Boris Ostrovsky
2016-07-11 15:38 ` Ian Jackson
2016-07-11 15:47 ` Ian Jackson
2016-07-11 16:07 ` Boris Ostrovsky
2016-07-08 10:10 ` Jan Beulich
2016-07-08 14:39 ` Boris Ostrovsky
2016-07-08 15:11 ` Jan Beulich
2016-07-08 16:14 ` Boris Ostrovsky
2016-08-01 10:09 ` Jan Beulich
2016-08-01 14:06 ` Boris Ostrovsky [this message]
2016-08-01 14:18 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 03/20] acpi/hvmloader: Initialize vm_gid data outside " Boris Ostrovsky
2016-07-08 10:18 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 04/20] acpi/hvmloader: Decide which SSDTs to install in hvmloader Boris Ostrovsky
2016-07-08 10:27 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 05/20] acpi/hvmloader: Move passthrough initialization from ACPI code Boris Ostrovsky
2016-07-08 10:46 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 06/20] acpi/hvmloader: Collect processor and NUMA info in hvmloader Boris Ostrovsky
2016-07-08 13:36 ` Jan Beulich
2016-07-08 15:08 ` Boris Ostrovsky
2016-07-08 15:14 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 07/20] acpi/hvmloader: Set TIS header address " Boris Ostrovsky
2016-07-08 13:38 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 08/20] acpi/hvmloader: Make providing IOAPIC in MADT optional Boris Ostrovsky
2016-07-08 13:41 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 09/20] acpi/hvmloader: Build WAET optionally Boris Ostrovsky
2016-07-08 13:42 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 10/20] acpi/hvmloader: Replace mem_alloc() and virt_to_phys() with memory ops Boris Ostrovsky
2016-07-08 13:58 ` Jan Beulich
2016-07-08 15:23 ` Boris Ostrovsky
2016-07-08 15:35 ` Jan Beulich
2016-07-08 16:19 ` Boris Ostrovsky
2016-07-19 9:11 ` Jan Beulich
2016-07-19 14:08 ` Boris Ostrovsky
2016-07-05 19:05 ` [PATCH v1 11/20] acpi/hvmloader: Translate all addresses when assigning addresses in ACPI tables Boris Ostrovsky
2016-07-08 14:31 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 12/20] acpi/hvmloader: Link ACPI object files directly Boris Ostrovsky
2016-07-08 14:51 ` Jan Beulich
2016-07-08 15:41 ` Boris Ostrovsky
2016-07-05 19:05 ` [PATCH v1 13/20] acpi/hvmloader: Include file/paths adjustments Boris Ostrovsky
2016-07-08 15:51 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 14/20] acpi: Move ACPI code to tools/libacpi Boris Ostrovsky
2016-08-03 16:00 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 15/20] x86: Add more checks verifying that PIT/PIC/IOAPIC are emulated Boris Ostrovsky
2016-08-03 16:04 ` Jan Beulich
2016-07-05 19:05 ` [PATCH v1 16/20] x86: Allow LAPIC-only emulation_flags for HVM guests Boris Ostrovsky
2016-08-03 16:11 ` Jan Beulich
2016-08-03 16:15 ` Andrew Cooper
2016-07-05 19:05 ` [PATCH v1 17/20] libacpi: Build DSDT for PVH guests Boris Ostrovsky
2016-07-05 19:05 ` [PATCH v1 18/20] libxl/acpi: Add ACPI e820 entry Boris Ostrovsky
2016-07-06 10:00 ` Julien Grall
2016-07-06 15:43 ` Boris Ostrovsky
2016-07-05 19:05 ` [PATCH v1 19/20] libxl/pvhv2: Include APIC page in MMIO hole for PVHv2 guests Boris Ostrovsky
2016-07-07 16:47 ` Wei Liu
2016-07-07 17:02 ` Boris Ostrovsky
2016-07-07 17:16 ` Wei Liu
2016-07-05 19:05 ` [PATCH v1 20/20] libxl/acpi: Build ACPI tables for HVMlite guests Boris Ostrovsky
2016-07-06 11:05 ` Julien Grall
2016-07-06 15:50 ` Boris Ostrovsky
2016-07-06 16:04 ` Julien Grall
2016-07-06 16:30 ` Boris Ostrovsky
2016-07-06 17:03 ` Julien Grall
2016-07-06 17:33 ` Boris Ostrovsky
2016-07-07 8:38 ` Jan Beulich
2016-07-07 15:08 ` Boris Ostrovsky
2016-07-07 15:12 ` Julien Grall
2016-07-07 15:24 ` Jan Beulich
2016-07-08 10:55 ` Wei Liu
2016-07-08 14:48 ` Boris Ostrovsky
2016-07-08 16:07 ` Wei Liu
2016-07-08 17:20 ` Boris Ostrovsky
2016-07-11 10:47 ` Wei Liu
2016-07-11 13:33 ` Boris Ostrovsky
2016-07-11 13:39 ` Julien Grall
2016-07-11 13:42 ` Wei Liu
2016-07-11 13:58 ` Julien Grall
2016-07-11 13:41 ` Wei Liu
2016-07-11 14:40 ` Boris Ostrovsky
2016-07-12 14:30 ` Wei Liu
2016-07-11 14:00 ` Anthony PERARD
2016-07-06 16:04 ` [PATCH v1 00/20] Make ACPI builder available to components other than hvmloader Roger Pau Monné
2016-07-06 16:32 ` Boris Ostrovsky
2016-07-07 8:35 ` Jan Beulich
2016-07-07 9:14 ` Julien Grall
2016-07-07 9:20 ` Jan Beulich
2016-07-07 9:29 ` Julien Grall
2016-07-07 15:04 ` Boris Ostrovsky
2016-07-07 15:10 ` Jan Beulich
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=40889253-ce31-eebe-cf78-8ad27dd31f82@oracle.com \
--to=boris.ostrovsky@oracle.com \
--cc=JBeulich@suse.com \
--cc=andrew.cooper3@citrix.com \
--cc=ian.jackson@eu.citrix.com \
--cc=julien.grall@arm.com \
--cc=roger.pau@citrix.com \
--cc=wei.liu2@citrix.com \
--cc=xen-devel@lists.xen.org \
--cc=zhaoshenglong@huawei.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).