Re: [PATCH 3.16 23/99] tty/ldsem: Wake up readers after timed out down_write()

From: Dmitry Safonov
Date: Tue Apr 02 2019 - 10:39:21 EST

On 4/2/19 3:32 PM, Ben Hutchings wrote:
> On Tue, 2019-04-02 at 15:22 +0100, Dmitry Safonov wrote:
>> Hi Ben,
>> JFI: I haven't checked your tree, but the same patches set contains fixes
>> that are more important to my mind (looking at 4.9 stable tree):
>> - "tty: Drop tty->count on tty_reopen() failure" commit fe3241679009
> You marked this as applicable to 4.6+. Is it actually applicable to
> older versions as well?

Oh, probably you're right - I'll need to look at the v3.16 source to
tell better. I think, I've marked the proper version at that time.

>> - "tty: Hold tty_ldisc_lock() during tty_reopen()" commit 83d817f41070
>> with follow-up fixup "tty: Don't hold ldisc lock in tty_reopen() if
>> ldisc present"
>> commit d3736d82e816
> [...]
> I will include these in a later update, unless you think they are
> really urgent and should be added to this one.

Well, I thought worth to mention those patches, but in reality haven't
checked if they are applicable to v3.16.
It's just I remember "tty: Hold tty_ldisc_lock() during tty_reopen()"
was the main fix in the set, as many people suffered from issue under
it, so I thought strange that only a side-patch (which can lead to soft
lockup, so probably also important) is ported. But I managed to forget
that the code has changes since v3.16.