linux-kernel - Re: [PATCH] scsi: mpt3sas: Fix memory allocation failure test in 'mpt3sas_base

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:   Mon, 7 Aug 2017 10:45:35 +0200
From:   Christophe JAILLET <christophe.jaillet@...adoo.fr>
To:     wharms@....de
Cc:     sathya.prakash@...adcom.com, chaitra.basappa@...adcom.com,
        suganath-prabu.subramani@...adcom.com, jejb@...ux.vnet.ibm.com,
        martin.petersen@...cle.com, MPT-FusionLinux.pdl@...adcom.com,
        linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel-janitors@...r.kernel.org
Subject: Re: [PATCH] scsi: mpt3sas: Fix memory allocation failure test in
 'mpt3sas_base_attach()'

Le 07/08/2017 à 10:25, walter harms a écrit :
>
> Am 07.08.2017 00:51, schrieb Christophe JAILLET:
>> In the lines above this test, 8 'kzalloc' are performed, but only 7 results
>> are tested.
>>
>> Add the missing one (i.e. '!ioc->port_enable_cmds.reply').
>>
>> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
>> ---
>>   drivers/scsi/mpt3sas/mpt3sas_base.c | 8 ++++----
>>   1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
>> index 1a5b6e40fb5c..8a44636ab0b5 100644
>> --- a/drivers/scsi/mpt3sas/mpt3sas_base.c
>> +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
>> @@ -5494,10 +5494,10 @@ mpt3sas_base_attach(struct MPT3SAS_ADAPTER *ioc)
>>   	ioc->ctl_cmds.status = MPT3_CMD_NOT_USED;
>>   	mutex_init(&ioc->ctl_cmds.mutex);
>>   
>> -	if (!ioc->base_cmds.reply || !ioc->transport_cmds.reply ||
>> -	    !ioc->scsih_cmds.reply || !ioc->tm_cmds.reply ||
>> -	    !ioc->config_cmds.reply || !ioc->ctl_cmds.reply ||
>> -	    !ioc->ctl_cmds.sense) {
>> +	if (!ioc->base_cmds.reply || !ioc->port_enable_cmds.reply ||
>> +	    !ioc->transport_cmds.reply || !ioc->scsih_cmds.reply ||
>> +	    !ioc->tm_cmds.reply || !ioc->config_cmds.reply ||
>> +	    !ioc->ctl_cmds.reply || !ioc->ctl_cmds.sense) {
>>   		r = -ENOMEM;
>>   		goto out_free_resources;
>>   	}
>
> obviously it is better to follow the pattern "malloc() , check".
Agreed, but it is also more verbose. Leavig it as-is, is IMHO, good enough.

> Even the programmer lost track.
>
> Bonus points if you malloc the buffers in one step.
Most of the allocation are 'kzalloc(ioc->reply_sz, GFP_KERNEL);', so a 
kcalloc could be used instead.
However, the 'kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_KERNEL);' breaks this 
logic and allocating all at once would lead to spaghetti code for no reason.

Moreover, I don't have any idea how big can be 'ioc->reply_sz', even if 
I guess it should be small.
So allocating all at once, could fail where several steps would work.

So I won't play for the bonus points :).

Best regards.
CJ

> just my 2 cents,
>
> re,
>   wh
>
>
>