From: "Wu, Jiaxin" <jiaxin.wu@intel.com>
To: Gerd Hoffmann <kraxel@redhat.com>, "Ni, Ray" <ray.ni@intel.com>
Cc: "devel@edk2.groups.io" <devel@edk2.groups.io>,
"Dong, Eric" <eric.dong@intel.com>,
"Zeng, Star" <star.zeng@intel.com>,
Laszlo Ersek <lersek@redhat.com>,
"Kumar, Rahul R" <rahul.r.kumar@intel.com>
Subject: Re: [PATCH v3 5/5] OvmfPkg/SmmCpuFeaturesLib: Skip SMBASE configuration
Date: Sun, 29 Jan 2023 05:47:00 +0000 [thread overview]
Message-ID: <MN0PR11MB6158F856BCC71EF65ABCD601FED29@MN0PR11MB6158.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20230119075303.nkyno36h25xscwkn@sirius.home.kraxel.org>
>
> * Why do you want do it in PEI instead? The cover letter claims this
> improves boot performance. I don't buy that. You are not skipping
> the relocation process, you are just doing it somewhere else.
I clarified why it can improves boot performance, it's not just do somewhere else, but also each cpus smbase can be programmed in parallel. See [PATCH v3 3/5] UefiCpuPkg/PiSmmCpuDxeSmm: Consume SMM Base Hob for SmBase info:
"> With SMM Base Hob, PiSmmCpuDxeSmm does not need the RSM
> instruction to reload the SMBASE register with the new allocated
> SMBASE each time when it exits SMM. SMBASE Register for each
See existing code smm cpu driver implementation:
//
// Hook return after RSM to set SMM re-based flag
//
SemaphoreHook (Index, &mRebased[Index]);
With above Semaphore hook, SMM CPU init for each processor must do one by one to avoid the CPUs over-writing each other's SMM Save State Area. As you mentioned, if system has thousands of cores, early smm init in parallel has the advantage over existing serial lines.
> processors have already been programmed and all SMBASE address
> have recorded in SMM Base Hob. So the same default SMBASE Address
> (0x30000) will not be used, thus the CPUs over-writing each
> other's SMM Save State Area will not happen in PiSmmCpuDxeSmm
> driver. This way makes the first SMI init can be executed in
> parallel and save boot time on multi-core system."
> * Where is the code?
See the design of existing function: SemaphoreHook (Index, &mRebased[Index]);
> * It is totally unclear whenever it is possible and/or useful to
> initialize SMM that way on OVMF.
>
Add the same handing logic code in OVMF is necessary to make sure it work. If someone produced such hob in OVMF platform, and OVMF also use the same Pi smm cpu driver, then it will be a problem. Changes in OVMF is make sure it runs into right way.
> take care,
> Gerd
next prev parent reply other threads:[~2023-01-29 5:47 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-18 9:56 [PATCH v3 0/5] Simplify SMM Relocation Process Wu, Jiaxin
2023-01-18 9:56 ` [PATCH v3 1/5] UefiCpuPkg/SmmBaseHob.h: Add SMM Base HOB Data Wu, Jiaxin
2023-01-18 11:19 ` Gerd Hoffmann
2023-01-18 15:06 ` Ni, Ray
2023-01-19 7:13 ` Gerd Hoffmann
2023-01-29 5:08 ` Wu, Jiaxin
2023-02-01 13:02 ` Gerd Hoffmann
2023-01-20 8:21 ` Laszlo Ersek
2023-01-29 5:24 ` Wu, Jiaxin
2023-02-01 13:14 ` Gerd Hoffmann
2023-02-02 0:44 ` Wu, Jiaxin
2023-02-02 3:54 ` [edk2-devel] " Ni, Ray
2023-02-02 3:52 ` Ni, Ray
2023-02-02 12:51 ` Gerd Hoffmann
2023-02-02 22:29 ` [edk2-devel] " Brian J. Johnson
2023-02-03 3:14 ` Ni, Ray
2023-02-03 7:54 ` Gerd Hoffmann
2023-02-03 13:22 ` Wu, Jiaxin
2023-02-03 13:31 ` Ni, Ray
2023-02-03 15:00 ` Gerd Hoffmann
2023-01-18 9:56 ` [PATCH v3 2/5] UefiCpuPkg/PiSmmCpuDxeSmm: Fix invalid InitializeMpSyncData call Wu, Jiaxin
2023-01-18 9:56 ` [PATCH v3 3/5] UefiCpuPkg/PiSmmCpuDxeSmm: Consume SMM Base Hob for SmBase info Wu, Jiaxin
2023-01-18 12:02 ` Gerd Hoffmann
2023-01-29 6:14 ` [edk2-devel] " Wu, Jiaxin
2023-01-18 9:56 ` [PATCH v3 4/5] UefiCpuPkg/SmmCpuFeaturesLib: Skip SMBASE configuration Wu, Jiaxin
2023-01-18 9:56 ` [PATCH v3 5/5] OvmfPkg/SmmCpuFeaturesLib: " Wu, Jiaxin
2023-01-18 12:19 ` Gerd Hoffmann
2023-01-18 14:37 ` Ni, Ray
2023-01-19 7:53 ` Gerd Hoffmann
2023-01-29 5:47 ` Wu, Jiaxin [this message]
2023-02-01 13:40 ` Gerd Hoffmann
2023-02-02 1:41 ` Wu, Jiaxin
2023-02-02 9:00 ` Gerd Hoffmann
2023-02-02 11:47 ` Laszlo Ersek
2023-02-02 12:24 ` Gerd Hoffmann
2023-02-03 3:05 ` Wu, Jiaxin
2023-02-03 2:47 ` Wu, Jiaxin
2023-02-03 3:45 ` Ni, Ray
2023-02-03 7:31 ` Gerd Hoffmann
2023-02-03 7:43 ` Ni, Ray
2023-02-03 8:49 ` Gerd Hoffmann
2023-02-03 11:18 ` Wu, Jiaxin
[not found] ` <173B5EAF72B992BD.14781@groups.io>
2023-02-03 8:59 ` [edk2-devel] " Wu, Jiaxin
2023-02-03 9:04 ` Gerd Hoffmann
2023-02-03 11:15 ` Wu, Jiaxin
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=MN0PR11MB6158F856BCC71EF65ABCD601FED29@MN0PR11MB6158.namprd11.prod.outlook.com \
--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