[PATCH] mtd: fix the wrong timeo for panic_nand_wait()

Huang Shijie shijie8 at gmail.com
Mon Jan 21 09:54:10 EST 2013


On Mon, Jan 21, 2013 at 10:27 PM, Matthieu CASTET
<matthieu.castet at parrot.com> wrote:
> Could you explain why the old code was wrong ?


sorry. this patch has a compiler error.


>
> With HZ=100
>
> timeo = jiffies + 2 and it should be working.
>

In logic, timeo should only be 2, just like it does in panic_nand_write().



thanks
Huang Shijie

>
> Huang Shijie a écrit :
>> In nand_wait(), the timeo for panic_nand_wait() is assigned with
>> wrong value(jiffies + some delay).
>>
>> This patch fixes it, and also uses the msecs_to_jiffies() to make the
>> code more readable.
>>
>> Signed-off-by: Huang Shijie <shijie8 at gmail.com>
>> ---
>>  drivers/mtd/nand/nand_base.c |    8 ++------
>>  1 files changed, 2 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
>> index 8323ac9..acdd508 100644
>> --- a/drivers/mtd/nand/nand_base.c
>> +++ b/drivers/mtd/nand/nand_base.c
>> @@ -825,14 +825,9 @@ static void panic_nand_wait(struct mtd_info *mtd, struct nand_chip *chip,
>>  static int nand_wait(struct mtd_info *mtd, struct nand_chip *chip)
>>  {
>>
>> -     unsigned long timeo = jiffies;
>> +     unsigned long timeo = (state == FL_ERASING) ? 400 : 20;
>>       int status, state = chip->state;
>>
>> -     if (state == FL_ERASING)
>> -             timeo += (HZ * 400) / 1000;
>> -     else
>> -             timeo += (HZ * 20) / 1000;
>> -
>>       led_trigger_event(nand_led_trigger, LED_FULL);
>>
>>       /*
>> @@ -849,6 +844,7 @@ static int nand_wait(struct mtd_info *mtd, struct nand_chip *chip)
>>       if (in_interrupt() || oops_in_progress)
>>               panic_nand_wait(mtd, chip, timeo);
>>       else {
>> +             timeo = jiffies + msecs_to_jiffies(timeo);
>>               while (time_before(jiffies, timeo)) {
>>                       if (chip->dev_ready) {
>>                               if (chip->dev_ready(mtd))



More information about the linux-mtd mailing list