On 7/7/20 5:32 AM, Andreas Grünbacher wrote:
>> @@ -2131,6 +2166,11 @@ ssize_t generic_file_buffered_read(struct kiocb *iocb,
>> }
>>
>> readpage:
>> + if (iocb->ki_flags & IOCB_NOWAIT) {
>> + unlock_page(page);
>> + put_page(page);
>> + goto would_block;
>> + }
>
> This hunk should have been part of "mm: allow read-ahead with
> IOCB_NOWAIT set" ...It probably should have... Oh well, it's been queued up multiple weeks at this point, not worth rebasing to move this hunk. -- Jens Axboe
