Re: [PATCH v2] drivers: net: xgene: Rewrite loop in xgene_enet_ecc_init()

[David Miller]

From: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
Date: Wed, 22 Oct 2014 21:50:06 +0200

> On Wed, Oct 22, 2014 at 9:34 PM, David Miller <davem@xxxxxxxxxxxxx> wrote:
>> From: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
>> Date: Wed, 22 Oct 2014 09:39:41 +0200
>>
>>> drivers/net/ethernet/apm/xgene/xgene_enet_sgmac.c: In function ʽxgene_enet_ecc_initʼ:
>>> drivers/net/ethernet/apm/xgene/xgene_enet_sgmac.c:126: warning: ʽdataʼ may be used uninitialized in this function
>>>
>>> Depending on the arbitrary value on the stack, the loop may terminate
>>> too early, and cause a bogus -ENODEV failure.
>>>
>>> Signed-off-by: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
>>> ---
>>> v2: Rewrite the loop instead of pre-initializing data.
>>
>> I hate to be a pest, but like the other patch of your's I think
>> a do { } while() works best here because the intent is clearly
>> to run the loop at least once, right?
> 
> I wanted to avoid checking for "data != ~0U" twice: once to abort the loop,
> and once to check if a timeout happened.

Hmmm:

	do {
		usleep_range(...);
		data = ...();
		if (data == ~0)
			return 0;
	} while (++i < 10);

	netdev_err(...);
	return -ENODEV;

Why would you have to check data twice?
N?§²ζμrΈ?yϊθ?Ψb²X¬ΆΗ§vΨ^?)ήΊ{.nΗ+?·??{±?κηzX§Ά?ʽά¨}©?²Ζ zΪ&j:+v?¨Ύ«?κηzZ+?Κ+zf£ʼ·h??§~?­?Ϋi?ϋΰzΉ?w?ʼΈ??¨θ­Ϊ&ʼ)ίʼf?ω^jΗ«y§m?α@A«aΆΪ?0Άμh?ε?i