Hi, Eric, I agree with your idea, I'll prepare another patch, thanks for remind this possible issue.
-- Best Regards, Feng Hong Application Processor Software Engnieer Marvell Technology (Shanghai) Ltd -----Original Message----- From: Eric W. Biederman [mailto:ebied...@xmission.com] Sent: 2012年9月19日 13:58 To: Feng Hong Cc: a...@linux-foundation.org; gorcu...@openvz.org; keesc...@chromium.org; serge.hal...@canonical.com; linux-kernel@vger.kernel.org Subject: Re: [PATCH] poweroff: fix bug in orderly_poweroff Feng Hong <hongf...@marvell.com> writes: > Hi, Eric > > 1. We are developing on an Android phone platform, we use thermal > framework to monitor the temperature, when the temperature above like > 110 degree, thermal framework will use orderly_shutdown to shutdown > phone, however, on Android platform there is no " /sbin/poweroff " cmd > ready . Then we want "fail ret" to trigger force shutdown (use > kernel_power_off), but always we get "suc ret" > 2. Here the caller just wait for "poweroff" userspace application, if > it block the called, then it's the "poweroff" problem itself > 3. As in the original orderly_shutdown design, we must get the right > "ret", if this ret is always "0", then it obey orderly_poweroff design > goal. Step 2: force shutdown is always useless code. That sounds like a clear case that we need to change it to UMH_WAIT_EXEC. Changing it to UMH_WAIT_PROC seems much more dangerous. Eric N�Р骒r��y����b�X�肚�v�^�)藓{.n�+�伐�{��赙zXФ�≤�}��财�z�&j:+v�����赙zZ+��+zf＂�h���~����i���z� �wア�?�ㄨ��&�)撷f��^j谦y�m��@A�a囤� 0鹅h���i
