linux-kernel - Re: [PATCH 10/11] pciehp: add support for bridge resource reallocation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Sun, 17 Jan 2010 14:35:35 -0800
From:	Yinghai Lu <yinghai@...nel.org>
To:	Jesse Barnes <jbarnes@...tuousgeek.org>
CC:	Bjorn Helgaas <bjorn.helgaas@...com>, Alex Chiang <achiang@...com>,
	Ingo Molnar <mingo@...e.hu>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Ivan Kokshaysky <ink@...assic.park.msu.ru>,
	Kenji Kaneshige <kaneshige.kenji@...fujitsu.com>,
	linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org
Subject: Re: [PATCH 10/11] pciehp: add support for bridge resource reallocation

On 01/17/2010 09:34 AM, Jesse Barnes wrote:
> On Sat, 16 Jan 2010 16:14:31 -0700
> Bjorn Helgaas <bjorn.helgaas@...com> wrote:
> 
>> On Sat, 2010-01-16 at 01:34 -0800, Yinghai Lu wrote:
>>> On 01/15/2010 08:59 PM, Alex Chiang wrote:
>>>> * Yinghai Lu <yinghai@...nel.org>:
>>>>> From: Kenji Kaneshige <kaneshige.kenji@...fujitsu.com>
>>>>>
>>>>> With this patch, pciehp driver try to clear PCI bridge resources
>>>>> to parent bridge (root port or switch downstream port) of the
>>>>> slot
>>>>>
>>>>> so we can shrink pci bridge resource for those port
>>>>>
>>>>> This feature is enabled when 'pciehp_realloc' option is
>>>>> specified.
>>>>>
>>>>> -v2: make it could be appiled after Yinghai patchset that touch
>>>>> pci bridge resource also remove poweron check, because
>>>>> pci_bridge_release_res will check child at first
>>>>
>>>> Same comment as my earlier patch. Why not just make this the
>>>> default behavior, instead of introducing yet another command line
>>>> parameter for users to guess at?
>>>
>>> it will break Eric's setup/
>>
>> I think this is a clue that we don't understand the problem well
>> enough yet.
>>
>> I'm opposed to adding kernel parameters for this sort of thing.  It is
>> unreasonable to expect users to figure out whether they need to use
>> this parameter or not.
>>
>> Special-case switches like this make it much harder to maintain the
>> code in the future.
> 
> Agreed.  Yinghai, what's wrong in the new reassignment code that
> causes Eric's setup to break?  Can we just fix that instead and enable
> reallocation by default?

no.

Eric's setup is:
hot plug one addon card into one hotplug slots and new addon cards have more hot plug slots.

SLOTA will have SLOT B, SLOT C, and SLOT D as children. and not devices in SLOTB/C/D at first.

with the pciehp_realloc is specified, the SLOTA will not get any device assigned, and later
if put devices in SLOTB/SLOTC/SLOTD, will not get resource for those devices.
and we can not go above to update SLOT A, because: some devices on SLOTB could already have driver loaded.

if you don't like pciehp_realloc at this point, please only apply 1-8 and 11.

Thanks

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