linux-kernel - Re: [PATCH 6/7] scsi: pm8001: use dev_and_phy_addr

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]

Message-ID: <569cb47c-af63-cf70-ae1a-4c5116dd4877@huawei.com>
Date:   Fri, 23 Sep 2022 11:30:49 +0100
From:   John Garry <john.garry@...wei.com>
To:     Jason Yan <yanaijie@...wei.com>, <martin.petersen@...cle.com>,
        <jejb@...ux.ibm.com>
CC:     <linux-scsi@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <hare@...e.com>, <hch@....de>, <bvanassche@....org>,
        <jinpu.wang@...ud.ionos.com>
Subject: Re: [PATCH 6/7] scsi: pm8001: use dev_and_phy_addr_same() instead of
 open coded

On 23/09/2022 11:13, Jason Yan wrote:
>>
>> Please explain why.
>>
>> I would assume that if those helpers were only used in libsas code 
>> (and not LLDDs) then they could be put in sas_internal.h and no need 
>> for export
>>
> 
> 
> Sorry, I did not make it clear. I mean we need to export 
> sas_find_attathed_phy() below. Not the sas address comparation helpers.

That seems fine to me.

About sas_find_attathed_phy() implementation,

 > +static inline int sas_find_attathed_phy(struct expander_device *ex_dev,
 > +                                       struct domain_device *dev)
 > +{
 > +       struct ex_phy *phy;
 > +       int phy_id;
 > +
 > +       for (phy_id = 0; phy_id < ex_dev->num_phys; phy_id++) {
 > +               phy = &ex_dev->ex_phy[phy_id];
 > +               if (SAS_ADDR(phy->attached_sas_addr)
 > +                       == SAS_ADDR(dev->sas_addr))
 > +                       return phy_id;
 > +       }
 > +
 > +       return ex_dev->num_phys;

Returning ex_dev->num_phys would seem ok, but then the LLDD needs to 
check that return against ex_dev->num_phys. It seems ok, but I'm still 
not 100% comfortable with that. Maybe returning -ENODEV may be better.

Or return boolean and pass phy_id as pointer to be filled in when 
returning true.

 > +}

Thanks,
John