linux-kernel - Re: [PATCH] mailbox: txdone_method shouldn't always be reset

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

Date:   Fri, 17 Nov 2017 09:51:40 +0000
From:   Sudeep Holla <sudeep.holla@....com>
To:     Bjorn Andersson <bjorn.andersson@...aro.org>,
        Jassi Brar <jassisinghbrar@...il.com>
Cc:     Sudeep Holla <sudeep.holla@....com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Alexey Klimov <klimov.linux@...il.com>
Subject: Re: [PATCH] mailbox: txdone_method shouldn't always be reset



On 17/11/17 07:04, Bjorn Andersson wrote:
> On Thu 16 Nov 22:47 PST 2017, Jassi Brar wrote:
>> On 16 Nov 2017 23:12, "Bjorn Andersson" <bjorn.andersson@...aro.org> wrote:
>> On Thu 16 Nov 09:06 PST 2017, Jassi Brar wrote:
>>> diff --git a/drivers/mailbox/mailbox.c b/drivers/mailbox/mailbox.c
>>> index 674b35f..95e480e 100644
>>> --- a/drivers/mailbox/mailbox.c
>>> +++ b/drivers/mailbox/mailbox.c
>>> @@ -124,7 +124,8 @@ static enum hrtimer_restart txdone_hrtimer(struct
>>> hrtimer *hrtimer)
>>>         for (i = 0; i < mbox->num_chans; i++) {
>>>                 struct mbox_chan *chan = &mbox->chans[i];
>>>
>>> -               if (chan->active_req && chan->cl) {
>>> +               if (chan->active_req && chan->cl &&
>>> +                   chan->txdone_method == TXDONE_BY_POLL) {
>>
>> The hrtimer code will crash before reaching this point if the channel
>> wasn't TXDONE_BY_POLL when it was created, so this part is not needed.
>>
>>
>> We have one timer for all channels of a controller. While this channel may
>> be run by ACK, some other might need to be POLLed. And we want to avoid
>> polling this channel.
>>
> 
> Oh, you're right.
> 
> But the fact that the timer function will poll channels that are
> "upgraded" to ACK is a separate issue.
> 

Ah right, I recall now that's the reason Alexey had patch introducing
timer per channel.

-- 
Regards,
Sudeep