Hi, Can somebody reproduce this bug? http://bugs.centos.org/view.php?id=4220 dd if=/dev/zero of=/dev/my_usb_stick_device_name bs=4096 count=10000 can not be stopped with ctrl-c and is not responding to signals (count=10000 can be modified to give you enough time to test) The bug has been closed with the following comment: "As said: dd is in an uninteruptable state when you do that. This is not a bug" This is wrong, because IMHO dd should receive the signal after it finishes the current read()/write() and returns to user space, but this is not happening. How can I reopen the bug? Thanks Adrian Buciuman
Adrian Buciuman wrote on Wed, 3 Mar 2010 13:32:53 +0200:> How can I reopen the bug?It is not a bug, maybe a feature request. *if* then this is to be reported upstream (= the dd developers), not in a distribution. I suppose it's specifically done this way, so it can't get interrupted accidentally. If you really want to kill it there are always ways to kill it by killing the process or shell. Kai -- Get your web at Conactive Internet Services: http://www.conactive.com
On 3/3/2010 5:32 AM, Adrian Buciuman wrote:> Hi, > > Can somebody reproduce this bug? > http://bugs.centos.org/view.php?id=4220 > > dd if=/dev/zero of=/dev/my_usb_stick_device_name bs=4096 count=10000 > can not be stopped with ctrl-c and is not responding to signals > > (count=10000 can be modified to give you enough time to test) > > The bug has been closed with the following comment: > "As said: dd is in an uninteruptable state when you do that. This is not a bug" > This is wrong, because IMHO dd should receive the signal after it > finishes the current read()/write() and returns to user space, but > this is not happening. > How can I reopen the bug?If you want to kill a process that doesn't want to be killed, just send it a -9. If that doesn't do it, it really is stuck in the device driver. -- Les Mikesell lesmikesell at gmail.com