Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1237366 > unrolled thread

QEMU fw_cfg DMA interface

Back to article view | Back to linux.kernel

Contents

  QEMU fw_cfg DMA interface Marc Marí <markmb@redhat.com> - 2015-10-01 14:20 +0200
    Re: [Qemu-devel] QEMU fw_cfg DMA interface Eric Blake <eblake@redhat.com> - 2015-10-01 18:10 +0200
      Re: [Qemu-devel] QEMU fw_cfg DMA interface Laszlo Ersek <lersek@redhat.com> - 2015-10-01 18:20 +0200
      Re: [Qemu-devel] QEMU fw_cfg DMA interface Laszlo Ersek <lersek@redhat.com> - 2015-10-01 18:20 +0200
        Re: [Qemu-devel] QEMU fw_cfg DMA interface Eric Blake <eblake@redhat.com> - 2015-10-01 18:30 +0200
          Re: [Qemu-devel] QEMU fw_cfg DMA interface Laszlo Ersek <lersek@redhat.com> - 2015-10-01 18:40 +0200
      Re: [Qemu-devel] QEMU fw_cfg DMA interface Eric Blake <eblake@redhat.com> - 2015-10-01 18:20 +0200

#1237366 — QEMU fw_cfg DMA interface

Implementation of the FW CFG DMA interface.

When running a Linux guest on top of QEMU, using the -kernel options, this
is the timing improvement for x86:

QEMU commit b2312c6 and SeaBIOS commit 423542e
QEMU startup time: .080
BIOS startup time: .060
Kernel setup time: .586
Total time: .726

QEMU with this patch series and SeaBIOS with this patch series
QEMU startup time: .080
BIOS startup time: .039
Kernel setup time: .005
Total time: .126

QEMU startup time is the time between the start and the first kvm_entry.
BIOS startup time is the time between the first kvm_entry and the start of
function do_boot, in SeaBIOS.
Kernel setup time is the time between the start of the function do_boot in
SeaBIOS and the jump to the Linux kernel.

As you can see, both the BIOS (because of ACPI tables and other configurations)
and the Linux kernel boot (because of the copy to memory) are greatly
improved with this new interface.

Also, this new interface is an addon to the old interface. Both interfaces
are compatible and interchangeable.

Changes from v1:
 - Take into account order of fields in the FWCfgDmaAccess structure
 - Check and change endianness of FWCfgDmaAccess fields
 - Change order of fields in the FWCfgDmaAccess structure
 - Add FW_CFG_DMA_CTL_SKIP feature for control field
 - Split FW_CFG_SIZE in QEMU
 - Make FW_CFG_ID a bitmap of features
 - Add 64 bit address support for the transfer. Trigger when writing the low
   address, and address is 0 by default and at the end of each transfer.
 - Align ports and addresses.
 - Preserve old fw_cfg_comb_valid behaviour in QEMU
 - Update documentation to reflect all these changes

Changes from v2:
 - Make IOports fw_cfg DMA region a different IO region.
 - Reuse everything for MMIO and IOport DMA regions
 - Make transfer status only based on control field
 - Use DMA helpers instead of direct map/unmap
 - Change ARM fw_cfg DMA address space
 - Change Linux boot process to match linuxboot.S
 - Add select capabilities in the FWCfgDmaAccess struct
 - Update documentation to reflect all these changes

Changes from v3:
 - Set properly fw_cfg DMA fields in ARM
 - Set fw_cfg DMA boot process properly (by Laszlo Ersek)
 - Add signature to fw_cfg DMA address field (by Kevin O'Connor)
 - Minor nitpicks
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1237544 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

[Multipart message — attachments visible in raw view] — view raw

[meta-comment]

On 10/01/2015 06:14 AM, Marc Marí wrote:
> Implementation of the FW CFG DMA interface.

The subject line is missing "v4" and "0/7". Also, the cover letter is
missing a diffstat.  That makes it harder to see from the cover letter
what the rest of the series is about.  'git format-patch/send-email
--cover-letter' does what you want; you can even 'git config
format.coverletter=auto' to always include a decent cover letter on any
multi-patch series.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[toc] | [prev] | [next] | [standalone]

#1237547 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

On 10/01/15 18:11, Eric Blake wrote:
> On 10/01/2015 10:03 AM, Eric Blake wrote:
>> [meta-comment]
>>
>> On 10/01/2015 06:14 AM, Marc Marí wrote:
>>> Implementation of the FW CFG DMA interface.
>>
>> The subject line is missing "v4" and "0/7". Also, the cover letter is
>> missing a diffstat.  That makes it harder to see from the cover letter
>> what the rest of the series is about.  'git format-patch/send-email
>> --cover-letter' does what you want; you can even 'git config
>> format.coverletter=auto' to always include a decent cover letter on any
>> multi-patch series.
> 
> Oh, I see - you sent a meta-cover letter (the one I replied to in this
> subthread), and then a patch series including a cover letter (the real
> 0/7, then 1/7 and friends in-reply-to the 0/7) as a child of the
> meta-cover.  It's still a bit awkward for tools that expect the 0/7 as
> the start of the thread,

Yep, the pattern I just described doesn't consider those tools. Is that
a bad problem? Maybe the pattern is not so clever then. :)

(I'm allowed to say bad things about it, because I "invented" it
"independently". :))

> and part of my confusion was caused by
> out-of-order mail delivery due to the nongnu.org mail server still
> recovering from its mail delays.
> 

Right, those delays are not helping.

Thanks
Laszlo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1237548 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

On 10/01/15 18:03, Eric Blake wrote:
> [meta-comment]
> 
> On 10/01/2015 06:14 AM, Marc Marí wrote:
>> Implementation of the FW CFG DMA interface.
> 
> The subject line is missing "v4" and "0/7". Also, the cover letter is
> missing a diffstat.  That makes it harder to see from the cover letter
> what the rest of the series is about.  'git format-patch/send-email
> --cover-letter' does what you want; you can even 'git config
> format.coverletter=auto' to always include a decent cover letter on any
> multi-patch series.
> 

This posting follows a little bit different pattern, one that I myself
follow when posting patches for two (or more) components that must work
in sync.

Usually, a top-level blurb is manually cross-posted to all relevant
mailing lists. Then, each separate patch series is posted only to the
relevant mailing list, with its own cover letter (as usual with git),
*in response* to the manually posted blurb.

This has the following benefits:

- in mailing list archives that organize messages into threads *across*
  mailing lists (like Gmane does, for example), the top-level manual
  blurb is a good "root" for referencing the entire posting.

- The same is true for personal mailboxes, if a recipient is explicitly
  CC'd on all of the messages.

Because the top level blurb is parent to several patch series, and those
child series can all have different version numbers (due to different
numbers of respinds), it is not always straightforward to assign a
version number to the top blurb.

Thanks
Laszlo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1237554 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

[Multipart message — attachments visible in raw view] — view raw

On 10/01/2015 10:17 AM, Laszlo Ersek wrote:
> On 10/01/15 18:03, Eric Blake wrote:
>> [meta-comment]
>>
>> On 10/01/2015 06:14 AM, Marc Marí wrote:
>>> Implementation of the FW CFG DMA interface.
>>
>> The subject line is missing "v4" and "0/7". Also, the cover letter is
>> missing a diffstat.  That makes it harder to see from the cover letter
>> what the rest of the series is about.  'git format-patch/send-email
>> --cover-letter' does what you want; you can even 'git config
>> format.coverletter=auto' to always include a decent cover letter on any
>> multi-patch series.
>>
> 
> This posting follows a little bit different pattern, one that I myself
> follow when posting patches for two (or more) components that must work
> in sync.

Ok, makes sense. Maybe the only additional suggestions would be to make
it more obvious in the subject line (put the text 'cross-post'
somewhere?) or have the first paragraph of the meta-cover be more
explicit that there are going to be multiple sub-threads, one per
project, where all subthreads must be applied to their corresponding
project for the overall feature to be complete?

[And maybe I should wait a few minutes for the full thread to appear in
my inbox, rather than immediately replying to the first mail while the
series is still incomplete due to mail delays...]

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[toc] | [prev] | [next] | [standalone]

#1237562 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

On 10/01/15 18:21, Eric Blake wrote:
> On 10/01/2015 10:17 AM, Laszlo Ersek wrote:
>> On 10/01/15 18:03, Eric Blake wrote:
>>> [meta-comment]
>>>
>>> On 10/01/2015 06:14 AM, Marc Marí wrote:
>>>> Implementation of the FW CFG DMA interface.
>>>
>>> The subject line is missing "v4" and "0/7". Also, the cover letter is
>>> missing a diffstat.  That makes it harder to see from the cover letter
>>> what the rest of the series is about.  'git format-patch/send-email
>>> --cover-letter' does what you want; you can even 'git config
>>> format.coverletter=auto' to always include a decent cover letter on any
>>> multi-patch series.
>>>
>>
>> This posting follows a little bit different pattern, one that I myself
>> follow when posting patches for two (or more) components that must work
>> in sync.
> 
> Ok, makes sense. Maybe the only additional suggestions would be to make
> it more obvious in the subject line (put the text 'cross-post'
> somewhere?) or have the first paragraph of the meta-cover be more
> explicit that there are going to be multiple sub-threads, one per
> project, where all subthreads must be applied to their corresponding
> project for the overall feature to be complete?

That's a good idea. I think prefixing the main blurb's subject with
[cross-post], and a "standard" first paragraph based on your above
suggestion, would be helpful.

> [And maybe I should wait a few minutes for the full thread to appear in
> my inbox, rather than immediately replying to the first mail while the
> series is still incomplete due to mail delays...]

I'm not patient; it would be unfair from me to expect others to be... :)

Laszlo

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1237552 — Re: [Qemu-devel] QEMU fw_cfg DMA interface

[Multipart message — attachments visible in raw view] — view raw

On 10/01/2015 10:03 AM, Eric Blake wrote:
> [meta-comment]
> 
> On 10/01/2015 06:14 AM, Marc Marí wrote:
>> Implementation of the FW CFG DMA interface.
> 
> The subject line is missing "v4" and "0/7". Also, the cover letter is
> missing a diffstat.  That makes it harder to see from the cover letter
> what the rest of the series is about.  'git format-patch/send-email
> --cover-letter' does what you want; you can even 'git config
> format.coverletter=auto' to always include a decent cover letter on any
> multi-patch series.

Oh, I see - you sent a meta-cover letter (the one I replied to in this
subthread), and then a patch series including a cover letter (the real
0/7, then 1/7 and friends in-reply-to the 0/7) as a child of the
meta-cover.  It's still a bit awkward for tools that expect the 0/7 as
the start of the thread, and part of my confusion was caused by
out-of-order mail delivery due to the nongnu.org mail server still
recovering from its mail delays.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web