vfs: vntblinit(): Raise default 'kern.maxvnodes' higher than 'kern.maxfiles'

Having 'kern.maxvnodes' higher than 'kern.maxfiles' mitigates a scenario
where some processes can eat up all vnodes in the system, causing
a deadlock, as long as the kernel itself does not create too many vnodes
without creating some file descriptor in some process' FD table.  A very
small percentage (~0.6%) of excess vnodes at infinity, coupled with
a large difference near the origin, should cover basic cases more than
enough.  Note however that this measure can be defeated, e.g., by using
nullfs mounts with non-trivial file hierarchies.

MFC after:      5 days
Sponsored by:   The FreeBSD Foundation
Differential Revision:  https://reviews.freebsd.org/D50314

(cherry picked from commit e9baf472a6)
(cherry picked from commit a6b05a35ce3c516cadfca49c310dffaabbe40440)

Approved by:    re (cperciva)
This commit is contained in:
Olivier Certner 2025-05-12 14:19:56 +02:00
parent 0b4d5c560a
commit 6b6d6448fe
No known key found for this signature in database
GPG key ID: 8CA13040971E2627

View file

@ -727,13 +727,12 @@ vntblinit(void *dummy __unused)
* KVA).
*
* Currently, on 64-bit platforms, 'desiredvnodes' is set to
* 'virtvnodes' up to a physical memory cutoff of ~1674MB, after which
* 'virtvnodes' up to a physical memory cutoff of ~1722MB, after which
* 'physvnodes' applies instead. With the current automatic tuning for
* 'maxfiles' (32 files/MB), 'desiredvnodes' becomes smaller than it at
* ~5136MB.
* 'maxfiles' (32 files/MB), 'desiredvnodes' is always greater than it.
*/
physvnodes = maxproc + pgtok(vm_cnt.v_page_count) / 64 +
3 * min(98304 * 16, pgtok(vm_cnt.v_page_count)) / 64;
physvnodes = maxproc + pgtok(vm_cnt.v_page_count) / 32 +
min(98304 * 16, pgtok(vm_cnt.v_page_count)) / 32;
virtvnodes = vm_kmem_size / (10 * (sizeof(struct vm_object) +
sizeof(struct vnode) + NC_SZ * ncsizefactor + NFS_NCLNODE_SZ));
desiredvnodes = min(physvnodes, virtvnodes);