Re: [PATCH] [14/18] BKL-removal: Add unlocked_fasync v2

> No goto if you use unlocked_fasync? Indeed. There was another problem in the patch too. Here's an updated patch that also fixes another latent bug. The whole f_flags still seems to be somewhat fragile because the checks tend to happen without any lock, but that has not changed to the previous

Re: [PATCH] [14/18] BKL-removal: Add unlocked_fasync

> +++ linux/fs/fcntl.c > @@ -240,11 +240,15 @@ static int setfl(int fd, struct file * f > > lock_kernel(); > if ((arg ^ filp->f_flags) & FASYNC) { > - if (filp->f_op && filp->f_op->fasync) { > + if (filp->f_op && filp->f_op->unlocked_fasync) > +

Re: [PATCH] [14/18] BKL-removal: Add unlocked_fasync

+++ linux/fs/fcntl.c @@ -240,11 +240,15 @@ static int setfl(int fd, struct file * f lock_kernel(); if ((arg ^ filp-f_flags) FASYNC) { - if (filp-f_op filp-f_op-fasync) { + if (filp-f_op filp-f_op-unlocked_fasync) +

Re: [PATCH] [14/18] BKL-removal: Add unlocked_fasync v2

No goto if you use unlocked_fasync? Indeed. There was another problem in the patch too. Here's an updated patch that also fixes another latent bug. The whole f_flags still seems to be somewhat fragile because the checks tend to happen without any lock, but that has not changed to the previous

Re: [PATCH] [14/18] BKL-removal: Add unlocked_fasync

> Add a new fops entry point to allow fasync without BKL. While it's arguably > unclear this entry point is called often enough for it really matters > it was still relatively easy to do. And there are far less async users > in the tree than ioctls so it's likely they can be all converted >

[PATCH] [14/18] BKL-removal: Add unlocked_fasync

Add a new fops entry point to allow fasync without BKL. While it's arguably unclear this entry point is called often enough for it really matters it was still relatively easy to do. And there are far less async users in the tree than ioctls so it's likely they can be all converted eventually and

[PATCH] [14/18] BKL-removal: Add unlocked_fasync

Add a new fops entry point to allow fasync without BKL. While it's arguably unclear this entry point is called often enough for it really matters it was still relatively easy to do. And there are far less async users in the tree than ioctls so it's likely they can be all converted eventually and