netdev - Re: [PATCH net-next v2] net: wwan: t7xx: fix GFP_KERNEL usage in spin

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:   Wed, 18 May 2022 12:39:21 +0800
From:   "Ziyang Xuan (William)" <william.xuanziyang@...wei.com>
To:     Loic Poulain <loic.poulain@...aro.org>
CC:     <chandrashekar.devegowda@...el.com>, <linuxwwan@...el.com>,
        <chiranjeevi.rapolu@...ux.intel.com>, <haijun.liu@...iatek.com>,
        <m.chetan.kumar@...ux.intel.com>,
        <ricardo.martinez@...ux.intel.com>, <ryazanov.s.a@...il.com>,
        <johannes@...solutions.net>, <davem@...emloft.net>,
        <edumazet@...gle.com>, <kuba@...nel.org>, <pabeni@...hat.com>,
        <netdev@...r.kernel.org>
Subject: Re: [PATCH net-next v2] net: wwan: t7xx: fix GFP_KERNEL usage in
 spin_lock context

> Hi Ziyang,
> 
> On Tue, 17 May 2022 at 08:30, Ziyang Xuan <william.xuanziyang@...wei.com> wrote:
>>
>> t7xx_cldma_clear_rxq() call t7xx_cldma_alloc_and_map_skb() in spin_lock
>> context, But __dev_alloc_skb() in t7xx_cldma_alloc_and_map_skb() uses
>> GFP_KERNEL, that will introduce scheduling factor in spin_lock context.
>>
>> Because t7xx_cldma_clear_rxq() is called after stopping CLDMA, so we can
>> remove the spin_lock from t7xx_cldma_clear_rxq().
>>
>> Fixes: 39d439047f1d ("net: wwan: t7xx: Add control DMA interface")
>> Signed-off-by: Ziyang Xuan <william.xuanziyang@...wei.com>
>> ---
> 
> You should normally indicate what changed in this v2.
> 
>>  drivers/net/wwan/t7xx/t7xx_hif_cldma.c | 7 ++++---
>>  1 file changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/net/wwan/t7xx/t7xx_hif_cldma.c b/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
>> index 46066dcd2607..7493285a9606 100644
>> --- a/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
>> +++ b/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
>> @@ -782,10 +782,12 @@ static int t7xx_cldma_clear_rxq(struct cldma_ctrl *md_ctrl, int qnum)
>>         struct cldma_queue *rxq = &md_ctrl->rxq[qnum];
>>         struct cldma_request *req;
>>         struct cldma_gpd *gpd;
>> -       unsigned long flags;
>>         int ret = 0;
>>
>> -       spin_lock_irqsave(&rxq->ring_lock, flags);
>> +       /* CLDMA has been stopped. There is not any CLDMA IRQ, holding
>> +        * ring_lock is not needed.
> 
> If it makes sense to explain why we don't need locking, the next
> sentence is not needed:

I want to remind the possible developer if he or she want to add spin_lock
here again in future, he or she should check whether there is a scheduling
factor or not here firstly.

> 
> 
>>  Thus we can use functions that may
>> +        * introduce scheduling.
>> +        */
>>         t7xx_cldma_q_reset(rxq);
>>         list_for_each_entry(req, &rxq->tr_ring->gpd_ring, entry) {
>>                 gpd = req->gpd;
>> @@ -808,7 +810,6 @@ static int t7xx_cldma_clear_rxq(struct cldma_ctrl *md_ctrl, int qnum)
>>
>>                 t7xx_cldma_gpd_set_data_ptr(req->gpd, req->mapped_buff);
>>         }
>> -       spin_unlock_irqrestore(&rxq->ring_lock, flags);
>>
>>         return ret;
>>  }
>> --
>> 2.25.1
>>
> .
>