All of lore.kernel.org
 help / color / mirror / Atom feed
From: Corey Minyard <tcminyard@gmail.com>
To: "Zheng, Lv" <lv.zheng@intel.com>
Cc: "Brown, Len" <len.brown@intel.com>,
	"Wysocki, Rafael J" <rafael.j.wysocki@intel.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Rafael J. Wysocki" <rjw@sisk.pl>,
	"linux-acpi@vger.kernel.org" <linux-acpi@vger.kernel.org>,
	"openipmi-developer@lists.sourceforge.net"
	<openipmi-developer@lists.sourceforge.net>,
	"Zhao, Yakui" <yakui.zhao@intel.com>
Subject: Re: [PATCH 03/13] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI transfers
Date: Thu, 25 Jul 2013 19:48:05 -0500	[thread overview]
Message-ID: <51F1C745.60800@acm.org> (raw)
In-Reply-To: <1AE640813FDE7649BE1B193DEA596E8802435A78@SHSMSX101.ccr.corp.intel.com>

On 07/25/2013 07:16 PM, Zheng, Lv wrote:
>>
>> If I understand this correctly, the problem would be if:
>>
>> rem_time = wait_for_completion_timeout(&tx_msg->tx_complete,
>>                                           IPMI_TIMEOUT);
>>
>> returns on a timeout, then checks msg_done and races with something setting
>> msg_done.  If that is the case, you would need the smp_rmb() before checking
>> msg_done.
>>
>> However, the timeout above is unnecessary.  You are using
>> ipmi_request_settime(), so you can set the timeout when the IPMI command
>> fails and returns a failure message.  The driver guarantees a return message
>> for each request.  Just remove the timeout from the completion, set the
>> timeout and retries in the ipmi request, and the completion should handle the
>> barrier issues.
> It's just difficult for me to determine retry count and timeout value, maybe retry=0, timeout=IPMI_TIMEOUT is OK.
> The code of the timeout completion is already there, I think the quick fix code should not introduce this logic.
> I'll add a new patch to apply your comment.

Since it is a local BMC, I doubt a retry is required.  That is probably 
fine.  Or you could set retry=1 and timeout=IPMI_TIMEOUT/2 if you wanted 
to be more sure, but I doubt it would make a difference.  The only time 
you really need to worry about retries is if you are resetting the BMC 
or it is being overloaded.

>
>> Plus, from a quick glance at the code, it doesn't look like it will properly handle a
>> situation where the timeout occurs and is handled then the response comes in
>> later.
> PATCH 07 fixed this issue.
> Here we just need the smp_rmb() or holding tx_msg_lock() around the acpi_format_ipmi_response().

If you apply the fix like I suggest, then the race goes away.  If 
there's no timeout and it just waits for the completion, things get a 
lot simpler.

>
> Thanks for commenting.

No problem, thanks for working on this.

-corey

------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk

WARNING: multiple messages have this Message-ID (diff)
From: Corey Minyard <tcminyard@gmail.com>
To: "Zheng, Lv" <lv.zheng@intel.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>,
	"Wysocki, Rafael J" <rafael.j.wysocki@intel.com>,
	"Brown, Len" <len.brown@intel.com>,
	"Zhao, Yakui" <yakui.zhao@intel.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-acpi@vger.kernel.org" <linux-acpi@vger.kernel.org>,
	"openipmi-developer@lists.sourceforge.net" 
	<openipmi-developer@lists.sourceforge.net>
Subject: Re: [PATCH 03/13] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI transfers
Date: Thu, 25 Jul 2013 19:48:05 -0500	[thread overview]
Message-ID: <51F1C745.60800@acm.org> (raw)
In-Reply-To: <1AE640813FDE7649BE1B193DEA596E8802435A78@SHSMSX101.ccr.corp.intel.com>

On 07/25/2013 07:16 PM, Zheng, Lv wrote:
>>
>> If I understand this correctly, the problem would be if:
>>
>> rem_time = wait_for_completion_timeout(&tx_msg->tx_complete,
>>                                           IPMI_TIMEOUT);
>>
>> returns on a timeout, then checks msg_done and races with something setting
>> msg_done.  If that is the case, you would need the smp_rmb() before checking
>> msg_done.
>>
>> However, the timeout above is unnecessary.  You are using
>> ipmi_request_settime(), so you can set the timeout when the IPMI command
>> fails and returns a failure message.  The driver guarantees a return message
>> for each request.  Just remove the timeout from the completion, set the
>> timeout and retries in the ipmi request, and the completion should handle the
>> barrier issues.
> It's just difficult for me to determine retry count and timeout value, maybe retry=0, timeout=IPMI_TIMEOUT is OK.
> The code of the timeout completion is already there, I think the quick fix code should not introduce this logic.
> I'll add a new patch to apply your comment.

Since it is a local BMC, I doubt a retry is required.  That is probably 
fine.  Or you could set retry=1 and timeout=IPMI_TIMEOUT/2 if you wanted 
to be more sure, but I doubt it would make a difference.  The only time 
you really need to worry about retries is if you are resetting the BMC 
or it is being overloaded.

>
>> Plus, from a quick glance at the code, it doesn't look like it will properly handle a
>> situation where the timeout occurs and is handled then the response comes in
>> later.
> PATCH 07 fixed this issue.
> Here we just need the smp_rmb() or holding tx_msg_lock() around the acpi_format_ipmi_response().

If you apply the fix like I suggest, then the race goes away.  If 
there's no timeout and it just waits for the completion, things get a 
lot simpler.

>
> Thanks for commenting.

No problem, thanks for working on this.

-corey

  reply	other threads:[~2013-07-26  0:48 UTC|newest]

Thread overview: 91+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <cover.1370652213.git.lv.zheng@intel.com>
2013-07-23  8:08 ` [PATCH 00/13] ACPI/IPMI: Fix several issues in the current codes Lv Zheng
2013-07-23  8:08   ` Lv Zheng
2013-07-23  8:08   ` [PATCH 01/13] ACPI/IPMI: Fix potential response buffer overflow Lv Zheng
2013-07-23  8:08     ` Lv Zheng
2013-07-23 14:54     ` Greg KH
2013-07-24  0:21       ` Zheng, Lv
2013-07-24  0:44       ` Zheng, Lv
2013-07-23  8:09   ` [PATCH 02/13] ACPI/IPMI: Fix atomic context requirement of ipmi_msg_handler() Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-23  8:09   ` [PATCH 03/13] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI transfers Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-24 23:38     ` Rafael J. Wysocki
2013-07-25  3:09       ` Zheng, Lv
2013-07-25  3:09         ` Zheng, Lv
2013-07-25 12:06         ` Rafael J. Wysocki
2013-07-25 18:12           ` Corey Minyard
2013-07-25 19:32             ` Rafael J. Wysocki
2013-07-26  0:18               ` Zheng, Lv
2013-07-26  0:18                 ` Zheng, Lv
2013-07-26  0:16             ` Zheng, Lv
2013-07-26  0:16               ` Zheng, Lv
2013-07-26  0:48               ` Corey Minyard [this message]
2013-07-26  0:48                 ` Corey Minyard
2013-07-26  1:30                 ` Zheng, Lv
2013-07-26  1:30                   ` Zheng, Lv
2013-07-26  0:09           ` Zheng, Lv
2013-07-26  0:09             ` Zheng, Lv
2013-07-23  8:09   ` [PATCH 04/13] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI user Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-25 21:59     ` Rafael J. Wysocki
2013-07-26  1:17       ` Zheng, Lv
2013-07-26  1:17         ` Zheng, Lv
2013-07-23  8:09   ` [PATCH 05/13] ACPI/IPMI: Fix issue caused by the per-device registration of the IPMI operation region handler Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-23  8:09   ` [PATCH 06/13] ACPI/IPMI: Add reference counting for ACPI operation region handlers Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-25 20:27     ` Rafael J. Wysocki
2013-07-26  0:47       ` Zheng, Lv
2013-07-26  0:47         ` Zheng, Lv
2013-07-26  8:09         ` Zheng, Lv
2013-07-26  8:09           ` Zheng, Lv
2013-07-26 14:00         ` Rafael J. Wysocki
2013-07-29  1:43           ` Zheng, Lv
2013-07-29  1:43             ` Zheng, Lv
2013-07-25 21:29     ` Rafael J. Wysocki
2013-07-26  1:54       ` Zheng, Lv
2013-07-26  1:54         ` Zheng, Lv
2013-07-26  8:15         ` Zheng, Lv
2013-07-26  8:15           ` Zheng, Lv
2013-07-26 14:49         ` Rafael J. Wysocki
2013-07-29  1:56           ` Zheng, Lv
2013-07-29  1:56             ` Zheng, Lv
2013-07-23  8:09   ` [PATCH 07/13] ACPI/IPMI: Add reference counting for ACPI IPMI transfers Lv Zheng
2013-07-23  8:09     ` Lv Zheng
2013-07-25 22:23     ` Rafael J. Wysocki
2013-07-26  1:21       ` Zheng, Lv
2013-07-26  1:21         ` Zheng, Lv
2013-07-26 13:41         ` Rafael J. Wysocki
2013-07-23  8:10   ` [PATCH 08/13] ACPI/IPMI: Cleanup several acpi_ipmi_device members Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-07-25 22:25     ` Rafael J. Wysocki
2013-07-26  1:25       ` Zheng, Lv
2013-07-26  1:25         ` Zheng, Lv
2013-07-26 13:38         ` Rafael J. Wysocki
2013-07-29  1:12           ` Zheng, Lv
2013-07-29  1:12             ` Zheng, Lv
2013-07-23  8:10   ` [PATCH 09/13] ACPI/IPMI: Cleanup some initialization codes Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-07-23  8:10   ` [PATCH 10/13] ACPI/IPMI: Cleanup some inclusion codes Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-07-23  8:10   ` [PATCH 11/13] ACPI/IPMI: Cleanup some Kconfig codes Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-07-23  8:10   ` [PATCH 12/13] Testing: Add module load/unload test suite Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-07-23  8:10   ` [PATCH 13/13] ACPI/IPMI: Add IPMI operation region test device driver Lv Zheng
2013-07-23  8:10     ` Lv Zheng
2013-09-13  5:13 ` [PATCH v2 00/12] ACPI/IPMI: Fix several issues in the current codes Lv Zheng
2013-09-13  5:13   ` [PATCH v2 01/12] ACPI/IPMI: Fix atomic context requirement of ipmi_msg_handler() Lv Zheng
2013-09-13  5:13   ` [PATCH v2 02/12] ACPI/IPMI: Fix potential response buffer overflow Lv Zheng
2013-09-13  5:13   ` [PATCH v2 03/12] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI transfers Lv Zheng
2013-09-13  5:13   ` [PATCH v2 04/12] ACPI/IPMI: Fix race caused by the timed out " Lv Zheng
2013-09-13  5:13   ` [PATCH v2 05/12] ACPI/IPMI: Fix race caused by the unprotected ACPI IPMI user Lv Zheng
2013-09-13  5:14   ` [PATCH v2 06/12] ACPI/IPMI: Fix issue caused by the per-device registration of the IPMI operation region handler Lv Zheng
2013-09-13  5:14   ` [PATCH v2 07/12] ACPI/IPMI: Add reference counting for ACPI IPMI transfers Lv Zheng
2013-09-13  5:14   ` [PATCH v2 08/12] ACPI/IPMI: Cleanup several acpi_ipmi_device members Lv Zheng
2013-09-13  5:14   ` [PATCH v2 09/12] ACPI/IPMI: Cleanup some initialization codes Lv Zheng
2013-09-13  5:14   ` [PATCH v2 10/12] ACPI/IPMI: Cleanup some inclusion codes Lv Zheng
2013-09-13  5:14   ` [PATCH v2 11/12] ACPI/IPMI: Cleanup some Kconfig codes Lv Zheng
2013-09-13  5:15   ` [PATCH v2 12/12] ACPI/IPMI: Cleanup coding styles Lv Zheng
2013-09-25 17:54   ` [PATCH v2 00/12] ACPI/IPMI: Fix several issues in the current codes Rafael J. Wysocki
2013-09-27  7:44     ` Zheng, Lv

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=51F1C745.60800@acm.org \
    --to=tcminyard@gmail.com \
    --cc=len.brown@intel.com \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lv.zheng@intel.com \
    --cc=minyard@acm.org \
    --cc=openipmi-developer@lists.sourceforge.net \
    --cc=rafael.j.wysocki@intel.com \
    --cc=rjw@sisk.pl \
    --cc=yakui.zhao@intel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.