On 2019/7/22 下午2:15, Nathan Dehnel wrote: > How do I get a longer backtrace?
If your dmesg output includes regular message other than the backtraces, it should be complete. If not, some dmesg may lost, you need to either check your system journal (if it catches all dmesg), or enlarge kernel message buffer (log_buf_len=<size> kernel parameter), or find a way to catch the message early enough to avoid unnecessary noise. Thanks, Qu > > On Mon, Jul 22, 2019, 12:02 AM Qu Wenruo <quwenruo.bt...@gmx.com > <mailto:quwenruo.bt...@gmx.com>> wrote: > > > > On 2019/7/22 下午12:25, Nathan Dehnel wrote: > > I'm still experiencing freezes with kernel 5.2. Here's a backtrace: > > > [snip] > > This time, there is no backtrace of btrfs functions at all, nor nfs > ones. > > Is the backtrace complete? If so, it would be something else causing the > problem. > > Thanks, > Qu > > > > > > > > On Sat, May 11, 2019 at 6:21 AM Qu Wenruo <quwenruo.bt...@gmx.com > <mailto:quwenruo.bt...@gmx.com>> wrote: > >> > >> > >>> [362108.291969] btrfs_tree_read_lock+0xbb/0xf1 > >>> [362108.291973] ? wait_woken+0x6d/0x6d > >>> [362108.291978] find_parent_nodes+0x91d/0x12b8 > >>> [362108.291985] ? btrfs_find_all_roots_safe+0x9c/0x107 > >>> [362108.291988] btrfs_find_all_roots_safe+0x9c/0x107 > >>> [362108.291992] btrfs_find_all_roots+0x57/0x75 > >>> [362108.291997] btrfs_qgroup_trace_extent_post+0x37/0x7c > >> > >> It's qgroup. > >> > >> We have upstream fix for it, 38e3eebff643 ("btrfs: honor > >> path->skip_locking in backref code"). > >> It's supported to be backported for all kernels after v4.14. > >> But I'm not sure if it's backported for your kernel. > >> > >> As you're gentoo user, it shouldn't be hard to check the kernel > source > >> to find if the fix is backported. > >> If not, feel free to backport for your kernel. > >> > >> Thanks, > >> Qu > >> > >> > >>> [362108.292002] btrfs_add_delayed_tree_ref+0x305/0x32b > >> >
signature.asc
Description: OpenPGP digital signature
