[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160129063656.GF4820@swordfish>
Date: Fri, 29 Jan 2016 15:37:24 +0900
From: Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>
To: Peter Hurley <peter@...leysoftware.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
Byungchul Park <byungchul.park@....com>,
akpm@...ux-foundation.org, mingo@...nel.org,
linux-kernel@...r.kernel.org, akinobu.mita@...il.com, jack@...e.cz,
torvalds@...ux-foundation.org
Subject: Re: [PATCH v4] lib/spinlock_debug.c: prevent a recursive cycle in
the debug code
On (01/29/16 15:16), Sergey Senozhatsky wrote:
>
> http://marc.info/?l=linux-kernel&m=144976121529901
>
hm... I don't like that patch. ->reset() loop must be done outside
of zap_locks(). we can have a printk() recursion in CPU1, but console
driver lock may be owned by CPU2 in driver's handle_IRQ(), for example.
stealing its lock CPU1 is not really good. in my kernels I do this from
panic() path only, where I know that things are already bad.
panic()->console_panic_mode()->{for_each_console()->reset(), zap_locks()}->console_trelock()->console_unlock().
-ss
Powered by blists - more mailing lists