linux-kernel - Re: [PATCH 5/5] wifi: mwifiex: use vmalloc

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: <c330d0f3-d3cd-4981-8bb5-3ff5761647dc@vivo.com>
Date: Wed, 13 Aug 2025 14:52:51 +0800
From: Qianfeng Rong <rongqianfeng@...o.com>
To: Dan Carpenter <dan.carpenter@...aro.org>
Cc: Brian Norris <briannorris@...omium.org>,
 Francesco Dolcini <francesco@...cini.it>,
 Johannes Berg <johannes.berg@...el.com>,
 Sascha Hauer <s.hauer@...gutronix.de>, Kalle Valo <kvalo@...nel.org>,
 Aditya Kumar Singh <quic_adisi@...cinc.com>,
 Rameshkumar Sundaram <quic_ramess@...cinc.com>,
 Roopni Devanathan <quic_rdevanat@...cinc.com>,
 "open list:MARVELL MWIFIEX WIRELESS DRIVER"
 <linux-wireless@...r.kernel.org>, open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 5/5] wifi: mwifiex: use vmalloc_array() to simplify code


在 2025/8/12 22:31, Dan Carpenter 写道:
> On Tue, Aug 12, 2025 at 10:13:43PM +0800, Qianfeng Rong wrote:
>> 在 2025/8/12 21:48, Dan Carpenter 写道:
>>> On Tue, Aug 12, 2025 at 09:32:18PM +0800, Qianfeng Rong wrote:
>>>> Remove array_size() calls and replace vmalloc() with vmalloc_array() to
>>>> simplify the code.
>>>>
>>>> Signed-off-by: Qianfeng Rong <rongqianfeng@...o.com>
>>>> ---
>>>>    drivers/net/wireless/marvell/mwifiex/cfg80211.c | 4 ++--
>>>>    1 file changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/net/wireless/marvell/mwifiex/cfg80211.c b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
>>>> index 3498743d5ec0..fb4183ff02a9 100644
>>>> --- a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
>>>> +++ b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
>>>> @@ -4673,8 +4673,8 @@ int mwifiex_init_channel_scan_gap(struct mwifiex_adapter *adapter)
>>>>    	 * additional active scan request for hidden SSIDs on passive channels.
>>>>    	 */
>>>>    	adapter->num_in_chan_stats = 2 * (n_channels_bg + n_channels_a);
>>>> -	adapter->chan_stats = vmalloc(array_size(sizeof(*adapter->chan_stats),
>>>> -						 adapter->num_in_chan_stats));
>>>> +	adapter->chan_stats = vmalloc_array(adapter->num_in_chan_stats,
>>>> +					    sizeof(*adapter->chan_stats));
>>> n_channels_bg is 14
>>> n_channels_a is either 0 or 31 depending on if we're using BAND_A.
>>> sizeof(*adapter->chan_stats) is 10.
>>>
>>> So we're either allocating 280 or 900 bytes, which is quite small.  We
>>> should just use kmalloc_array() instead of vmalloc_array().
>> Switching to kmalloc_array() requires changing vfree() to kfree(),
>> but I couldn't locate the memory release code.
> It's not hard to locate with a bit of looking.  ;)
Yes, I did find it. ;)
>>    This modification
>> likely requires deep familiarity with the codebase.  Also, for
>> variable-sized allocations, kvmalloc_array() is preferable, but
>> requires kvfree() for proper memory release.
> Don't use kvmalloc_array().  The rules are:
> 1) small amounts of memory: kmalloc()
> 2) possibly large ammounts of memory up to 2GB: kvmalloc()
> 3) definitely large ammounts of memory or larger than 2GB: vmalloc()
Allocations under 8 pages should use kmalloc() for efficiency.  For
allocations exceeding 8 pages, kvmalloc() or vmalloc() is preferable
(of course, non-contiguous physical memory can be used) since
kernel memory management defines PAGE_ALLOC_COSTLY_ORDER = 3
(indicating 8 pages), beyond which contiguous physical memory
allocation becomes unreliable and costly.
>
> There are also places where the memory needs to be contiguous and for
> those situations kvmalloc() and vmalloc() can't be used.
>
> Here we're allocating less than a PAGE so the appropriate thing is
> kmalloc_array().
>
> I'm looking at this some more and these statistics are updated in
> mwifiex_update_chan_statistics() and not necessarily all of them
> are updated.  But they're used in mwifiex_cfg80211_dump_survey()
> when potentially the memory is still uninitialized.
>
> I really think we should zero this memory, so lets allocate it
> with kcalloc().
I agree with you, I will try to do this in the next version.
>
> regards,
> dan carpenter
>