[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAHk-=wjBYuEf8XCBCNvf0vHf0RatNR308k+CO05UbwYnyTEL=w@mail.gmail.com>
Date: Sun, 22 Sep 2019 11:24:15 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Jessica Yu <jeyu@...nel.org>
Cc: Matthias Maennich <maennich@...gle.com>,
Martijn Coenen <maco@...roid.com>,
Masahiro Yamada <yamada.masahiro@...ionext.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [GIT PULL] Modules updates for v5.4
On Sun, Sep 22, 2019 at 8:28 AM Jessica Yu <jeyu@...nel.org> wrote:
>
> The main bulk of this pull request introduces a new exported symbol
> namespaces feature.
Hmm.
So I merged this, and felt that I tested my merge resolution fairly
well, and then pushed it out.
But during more extensive testing for special cases, I notice that
MODULE_IMPORT_NS() does not work at all when modules are disabled.
In particular, it causes errors like this:
drivers/usb/storage/uas.c:1222:1: warning: data definition has no
type or storage class
1222 | MODULE_IMPORT_NS(USB_STORAGE);
| ^~~~~~~~~~~~~~~~
because without modules, MODULE_IMPORT_NS() doesn't get #define'd at all.
This is not a merge error - it's an error in the original.
Tssk. I hate finding stupid bugs like this myself. Particularly when I
find them quickly from a simple "what if" test.
Yes, some of it goes to just bad luck with linux-next, but some of
this is simply people not thinking about different cases.
Tssk tssk.
Linus
Powered by blists - more mailing lists