Re: git: ec3c225711ae - main - vfs: call vn_truncate_locked from kern_truncate

weird. i'm on it.

On 5/20/22, Dmitry Chagin <dchagin@freebsd.org> wrote:
> On Fri, May 20, 2022 at 05:25:59PM +0000, Mateusz Guzik wrote:
>> The branch main has been updated by mjg:
>>
>> URL:
>> https://cgit.FreeBSD.org/src/commit/?id=ec3c225711ae7d5b8e71ed036aada09ef91149ec
>>
>> commit ec3c225711ae7d5b8e71ed036aada09ef91149ec
>> Author:     Mateusz Guzik <mjg@FreeBSD.org>
>> AuthorDate: 2022-05-16 01:28:22 +0000
>> Commit:     Mateusz Guzik <mjg@FreeBSD.org>
>> CommitDate: 2022-05-20 17:25:51 +0000
>>
>>     vfs: call vn_truncate_locked from kern_truncate
>>
>>     This fixes a bug where the syscall would not bump writecount.
>>
>>     PR:     263999
>> ---
>>  sys/kern/vfs_syscalls.c | 22 ++++++++++++----------
>>  1 file changed, 12 insertions(+), 10 deletions(-)
>>
>> diff --git a/sys/kern/vfs_syscalls.c b/sys/kern/vfs_syscalls.c
>> index 86ef3fc7a095..600299fbe6eb 100644
>> --- a/sys/kern/vfs_syscalls.c
>> +++ b/sys/kern/vfs_syscalls.c
>> @@ -3460,7 +3460,6 @@ kern_truncate(struct thread *td, const char *path,
>> enum uio_seg pathseg,
>>  	struct mount *mp;
>>  	struct vnode *vp;
>>  	void *rl_cookie;
>> -	struct vattr vattr;
>>  	struct nameidata nd;
>>  	int error;
>>
>> @@ -3480,18 +3479,21 @@ retry:
>>  		return (error);
>>  	}
>>  	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
>> -	if (vp->v_type == VDIR)
>> +	if (vp->v_type == VDIR) {
>>  		error = EISDIR;
>> -#ifdef MAC
>> -	else if ((error = mac_vnode_check_write(td->td_ucred, NOCRED, vp))) {
>> +		goto out;
>>  	}
>> +#ifdef MAC
>> +	error = mac_vnode_check_write(td->td_ucred, NOCRED, vp);
>> +	if (error != 0)
>> +		goto out;
>>  #endif
>> -	else if ((error = vn_writechk(vp)) == 0 &&
>> -	    (error = VOP_ACCESS(vp, VWRITE, td->td_ucred, td)) == 0) {
>> -		VATTR_NULL(&vattr);
>> -		vattr.va_size = length;
>> -		error = VOP_SETATTR(vp, &vattr, td->td_ucred);
>> -	}
>> +	error = VOP_ACCESS(vp, VWRITE, td->td_ucred, td);
>> +	if (error != 0)
>> +		goto out;
>> +
>> +	error = vn_truncate_locked(vp, 0, false, td->td_ucred);
>> +out:
>>  	VOP_UNLOCK(vp);
>>  	vn_finished_write(mp);
>>  	vn_rangelock_unlock(vp, rl_cookie);
>
> seems this broke truncate, test attached.
>
>
>


-- 
Mateusz Guzik <mjguzik gmail.com>