Hi Fabio, On 4/3/22 15:37, Fabio M. De Francesco wrote:
> > drivers/staging/r8188eu/core/rtw_pwrctrl.c:379 > > if (pwrpriv->ps_processing) {> while (pwrpriv->ps_processing && rtw_get_passing_time_ms(start) <= 3000) > msleep(10); > }> Hm, just wondering, shouldn't we annotate load from pwrpriv->ps_processing with READ_ONCE() inside while loop? IIUC compiler might want to cache first load into register and we will stuck here forever.You're right. This can be cached. In situations like these one should use barriers or other API that use barriers implicitly (completions, for example).
Not sure about completions, since they may sleep.Also, don't think that barriers are needed here, since this code just waiting for observing value 1. Might be barrier will slightly speed up waiting thread, but will also slow down other thread
In this list, a couple of weeks ago, I suggested someone to use them. He/she re-submitted his/her patches with wait_for_completion() / complete(), I placed my Reviewed-by tag and Greg was happy to accept it and merge. Not sure about READ_ONCE(), because, as far as I know, it is used only to prevent load tearing. But I might be wrong here. Fabio P.S.: Is Paul's book still un-read? Unfortunately, I am not yet done :(
If I could read it 24/7 I will be the happiest person ever :) As you might expect I am not yet done as well
With regards, Pavel Skripkin
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature
