linux-kernel - Re: [PATCH v2] eventfd: convert to f_op->read

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 1 May 2020 11:49:07 -0600
From:   Jens Axboe <axboe@...nel.dk>
To:     Al Viro <viro@...iv.linux.org.uk>
Cc:     linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] eventfd: convert to f_op->read_iter()

On 5/1/20 11:43 AM, Al Viro wrote:
> On Fri, May 01, 2020 at 11:18:05AM -0600, Jens Axboe wrote:
> 
>> -	if (res > 0 && put_user(ucnt, (__u64 __user *)buf))
>> +	if (res > 0 && copy_to_iter(&ucnt, res, iov) < res)
> 
> *whoa*
> 
> It is correct, but only because here res > 0 <=> res == 8.
> And that's not trivial at the first glance.
> 
> Please, turn that into something like

Looks good to me, just one minor edit:

> 	if (iov_iter_count(to) < sizeof(ucnt))
> 		return -EINVAL;
> 	spin_lock_irq(&ctx->wqh.lock);
> 	if (!ctx->count) {
> 		if (unlikely(file->f_flags & O_NONBLOCK) {
> 			spin_unlock_irq(&ctx->wqh.lock)
> 			return -EAGAIN;
> 		}
> 		__add_wait_queue(&ctx->wqh, &wait);
> 		for (;;) {
> 			set_current_state(TASK_INTERRUPTIBLE);
> 			if (ctx->count)
> 				break;
> 			if (signal_pending(current)) {
> 				spin_unlock_irq(&ctx->wqh.lock)
> 				return -ERESTARTSYS;
> 			}

We need to remove waitq and re-set task state here. I'll run a sanity
check on that and send out a v3.

-- 
Jens Axboe