Re: [PATCH 01/10] firewire-net: Use kmalloc_array() in fwnet_broadcast_start()

From: SF Markus Elfring
Date: Sat Sep 24 2016 - 11:30:02 EST


>> @@ -1103,8 +1103,7 @@ static int fwnet_broadcast_start(struct fwnet_device *dev)
>>
>> max_receive = 1U << (dev->card->max_receive + 1);
>> num_packets = (FWNET_ISO_PAGE_COUNT * PAGE_SIZE) / max_receive;
>> -
>> - ptrptr = kmalloc(sizeof(void *) * num_packets, GFP_KERNEL);
>> + ptrptr = kmalloc_array(num_packets, sizeof(*ptrptr), GFP_KERNEL);
>> if (!ptrptr) {
>> retval = -ENOMEM;
>> goto failed;
>
> Coccinelle enabled you to determine that kmalloc_array /could/ be used here.

A script for the semantic patch language pointed hundreds of source files out
with such software update opportunities.


> But whether it /should/ be used here is another question, and it is
> not addressed in your changelog.

I can expand the corresponding description when it will be desired.


> (You state that there is an "issue" but do not explain.)

Do you prefer an other wording for such an update candidate?


> kmalloc_array is a kmalloc wrapper which adds an inline check for integer
> overflow. So, can sizeof(void *) * num_packets ever overflow size_t?
>
> If yes,

Is there a probability that the calculated number of packets will become
too big for the preferred system limits anyhow?


> do we want a runtime check here (which kmalloc_array provides),

Did you notice the information from the commit "mm: faster kmalloc_array(), kcalloc()"
(from 2016-07-26) already?
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/?id=91c6a05f72a996bee5133e76374ab3ad7d3b9b72


> or do we want a compile-time check?

I guess that some software developers and subsystem maintainers are looking
for a bit more clarification around involved design dependencies.


> If no,
> then the remaining benefit of the patch is that it is more obvious
> to the reader that dev->broadcast_rcv_buffer_ptrs is an array,

How do you value such a kind of source code annotation?


> but possibly at the cost of superfluous code.

How do you think about to care for a bit more consistent use of Linux programming interfaces?


> Is gcc's optimizer able to resolve kmalloc_array's check at compile time
> as always false, such that the superfluous code is eliminated as dead code?

Which versions of compiler implementations would you like to check further?


> I believe I know answers to this but prefer to hear what you as the patch
> author think about it.

I presented another update suggestion also for this software module as a result
from a general source code search pattern.
The corresponding change acceptance varies and is evolving as usual.

Regards,
Markus