public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Ming Huang <ming.huang@linaro.org>
To: Leif Lindholm <leif.lindholm@linaro.org>
Cc: linaro-uefi@lists.linaro.org, edk2-devel@lists.01.org,
	graeme.gregory@linaro.org, ard.biesheuvel@linaro.org,
	michael.d.kinney@intel.com, lersek@redhat.com,
	wanghuiqiang@huawei.com, huangming23@huawei.com,
	zhangjinsong2@huawei.com, huangdaode@hisilicon.com,
	john.garry@huawei.com, xinliang.liu@linaro.org,
	zhangfeng56@huawei.com
Subject: Re: [PATCH edk2-non-osi v1 2/7] Hisilicon/D0x: Rename StartupAp() function
Date: Tue, 12 Feb 2019 20:07:52 +0800	[thread overview]
Message-ID: <51a23852-cfb8-a6c0-9a0d-21c39143d506@linaro.org> (raw)
In-Reply-To: <20190212104459.lifteascit2anj6h@bivouac.eciton.net>



On 2/12/2019 6:44 PM, Leif Lindholm wrote:
> On Tue, Feb 12, 2019 at 04:05:50PM +0800, Ming Huang wrote:
>> On 2/12/2019 5:36 AM, Leif Lindholm wrote:
>>> On Fri, Feb 01, 2019 at 10:25:02PM +0800, Ming Huang wrote:
>>>> As suggestion of community, 'AP' is a bit unfortunate to use in EDK2
>>>> context. PI specifies 'BSP' for Boot-strap Processor, as the one
>>>> executing all of the EDK2 code. It then uses 'AP' to refer to
>>>> Additional Processors, which can be assigned tasks using the
>>>> EFI_MP_SERVICES_PROTOCOL. In a TianoCore context, this should be
>>>> 'BSP'. So, Rename StartupAp() to StartUpBSP.
>>>
>>> Please add a comment somewhere that this applies to D0x
>>> PlatformSysCtrlLib.
>>
>> ok
>>
>>>> Contributed-under: TianoCore Contribution Agreement 1.1
>>>> Signed-off-by: Ming Huang <ming.huang@linaro.org>
>>>> ---
>>>>  Silicon/Hisilicon/Hi1610/Library/PlatformSysCtrlLibHi1610/PlatformSysCtrlLibHi1610.lib | Bin 297590 -> 229128 bytes
>>>>  Silicon/Hisilicon/Hi1616/Library/PlatformSysCtrlLibHi1616/PlatformSysCtrlLibHi1616.lib | Bin 344310 -> 275312 bytes
>>>>  Silicon/Hisilicon/Hi1620/Library/PlatformSysCtrlLibHi1620/PlatformSysCtrlLibHi1620.lib | Bin 356032 -> 375916 bytes
>>>>  3 files changed, 0 insertions(+), 0 deletions(-)
>>>
>>> These are substantial changes in image size from only changing the
>>> name of a function. So I'll have a little look around.
>>>
>>> 1610 version appears to have switched from building with GCC49_RELEASE
>>> to GCC48_RELEASE.
>>> 1616 and 1620 versions seem to have used GCC48_RELEASE all along.
>>>
>>> I definitely see additional renamed functions in these libraries too.
>>>
>>> Please have an inventory and determine what may be affecting image sizes.
>>>
>>> Also, I *beg* you - please upgrade from "GNU C 4.8.3 20131202 (prerelease)".
>>
>> We have plan to upgrage gcc to 7.3, but our build server is share for all ARM
>> project, so need discuss with other project groups, it may be not enough time
>> for 19.02.
> 
> Oh, we're too late in the game to change for this release.
> But you are using ancient toolchains with poor code generation and
> quite likely known bugs.
> And this has been the state for quite some time.
> If that can change for 19.06, that's good enough.

Ok, I will upgrade gcc from 4.8 for 19.06.

> 
>> For D05/D03 libraries, just remove 2 functions from OemMiscLib which used
>> by PlatformSysCtrlLib. Does edk2 version effect the libraries size?
>> old edk2 base on: 2017-0904
>> now edk2 base on: 2018-0801
> 
> Well, changing edk2 version will mean that command line options to
> compiler and linker may change. So certainly some change can be seen.
> But when the changes are this big, I suspect something else has been
> going on.

I also think it is strange for big size change.

> 
>> For D06 library, we use the same source code to support all Hi1620 projects,
>> include product projects,so there are some modify for this, like support
>> 3 sockets, 4 sockets and remove some useless functions.
> 
> So please reword the subject line of this commit to explain it is an
> overall update of PlatformSysCtrlLib - including which bits are dropped.
> 
> And I think this makes a good argument for moving the header files for
> binary-only libraries from edk2-platforms to edk2-non-osi.
> If you do that in a separate patch before this one, you won't need to
> include as much detail in the commit message as you will otherwise.

Do youe mean move PlatformSysCtrlLib.h, OemAddressMapLib.h and LpcLib.h to edk2-non-osi?

> 
> Regards,
> 
> Leif
> 
>> Thanks.
>>
>>>
>>> /
>>>     Leif
>>>
>>>>
>>>> diff --git a/Silicon/Hisilicon/Hi1610/Library/PlatformSysCtrlLibHi1610/PlatformSysCtrlLibHi1610.lib b/Silicon/Hisilicon/Hi1610/Library/PlatformSysCtrlLibHi1610/PlatformSysCtrlLibHi1610.lib
>>>> index 68be770..4c63a26 100644
>>>> Binary files a/Silicon/Hisilicon/Hi1610/Library/PlatformSysCtrlLibHi1610/PlatformSysCtrlLibHi1610.lib and b/Silicon/Hisilicon/Hi1610/Library/PlatformSysCtrlLibHi1610/PlatformSysCtrlLibHi1610.lib differ
>>>> diff --git a/Silicon/Hisilicon/Hi1616/Library/PlatformSysCtrlLibHi1616/PlatformSysCtrlLibHi1616.lib b/Silicon/Hisilicon/Hi1616/Library/PlatformSysCtrlLibHi1616/PlatformSysCtrlLibHi1616.lib
>>>> index b3cc88e..cb2c652 100644
>>>> Binary files a/Silicon/Hisilicon/Hi1616/Library/PlatformSysCtrlLibHi1616/PlatformSysCtrlLibHi1616.lib and b/Silicon/Hisilicon/Hi1616/Library/PlatformSysCtrlLibHi1616/PlatformSysCtrlLibHi1616.lib differ
>>>> diff --git a/Silicon/Hisilicon/Hi1620/Library/PlatformSysCtrlLibHi1620/PlatformSysCtrlLibHi1620.lib b/Silicon/Hisilicon/Hi1620/Library/PlatformSysCtrlLibHi1620/PlatformSysCtrlLibHi1620.lib
>>>> index 50d453a..d643f7b 100644
>>>> Binary files a/Silicon/Hisilicon/Hi1620/Library/PlatformSysCtrlLibHi1620/PlatformSysCtrlLibHi1620.lib and b/Silicon/Hisilicon/Hi1620/Library/PlatformSysCtrlLibHi1620/PlatformSysCtrlLibHi1620.lib differ
>>>> -- 
>>>> 2.9.5
>>>>


  reply	other threads:[~2019-02-12 12:08 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-01 14:25 [PATCH edk2-non-osi v1 0/7] Upload D0x binary modules Ming Huang
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 1/7] Hisilicon/D06: Optimize SAS driver for reducing boot time Ming Huang
2019-02-12 15:20   ` Leif Lindholm
2019-02-12 15:34     ` Ming Huang
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 2/7] Hisilicon/D0x: Rename StartupAp() function Ming Huang
2019-02-11 21:36   ` Leif Lindholm
2019-02-12  8:05     ` Ming Huang
2019-02-12 10:44       ` Leif Lindholm
2019-02-12 12:07         ` Ming Huang [this message]
2019-02-12 12:17           ` Leif Lindholm
2019-02-12 12:20             ` Ming Huang
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 3/7] Hisilicon/D06: Update Mbigen and gic RAS register Ming Huang
2019-02-11 21:38   ` Leif Lindholm
2019-02-12 11:42     ` Ming Huang
2019-02-12 12:06       ` Leif Lindholm
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 4/7] Hisilicon/D06: Support PCIe local RAS Ming Huang
2019-02-12 15:20   ` Leif Lindholm
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 5/7] Hisilicon/D06: Use new flash layout Ming Huang
2019-02-12 15:26   ` Leif Lindholm
2019-02-13 12:09     ` Ming Huang
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 6/7] Hisilicon/D06: Fix numa node wrong issue Ming Huang
2019-02-11 14:48   ` Leif Lindholm
2019-02-12 12:17     ` Ming Huang
2019-02-01 14:25 ` [PATCH edk2-non-osi v1 7/7] Hisilicon/D06: Add Setup Item "Support DPC" Ming Huang
2019-02-12 15:27   ` Leif Lindholm

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-list 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=51a23852-cfb8-a6c0-9a0d-21c39143d506@linaro.org \
    --to=devel@edk2.groups.io \
    /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