Dear Developers,
I hope this message finds you well. I've recently been examining an NFS
patch (linux-stable-v4.19.298 commit ID: fecb9d534dee) that I find
intriguing, and I was hoping to gain some clarity on a few of its
details.
The commit message states, "The NFS read code can trigger writeback
while holding the page lock. If an error then triggers a call to
nfs_write_error_remove_page(), we can deadlock." In the context as I
understand it, I'm having trouble conceptualizing the specific scenario
where this could lead to a deadlock.
I've scrutinized the relevant functions and code, including nfs_readpage
and generic_error_remove_page(), but I can't seem to pinpoint the exact
scenario where lock contention could lead to a deadlock.
I would greatly appreciate if you could help me understand this issue
and how this patch addresses it. What were the specific workloads or
system states you encountered in this process? Was this problem
discovered under specific hardware or configurations? Or does the
problem trigger under a certain sequence of operations?
Thank you for your assistance in advance, and I look forward to your
response.
Link:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-4.19.y&id=fecb9d534deed3680bfaff0b86ac83470946b710
Best regards,
ZhaoLong Wang
