From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marcelo Tosatti Subject: Re: more about serial ports: do they even work? Date: Wed, 4 Feb 2009 06:09:42 -0200 Message-ID: <20090204080942.GA3640@amt.cnet> References: <497E1B15.2090908@msgid.tls.msk.ru> <497E1F7D.90300@msgid.tls.msk.ru> <49876547.1080904@cisco.com> <49876678.4090808@msgid.tls.msk.ru> <4987674E.9050201@cisco.com> <4987FCBA.7020104@msgid.tls.msk.ru> <4988032A.5010905@ntlworld.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Michael Tokarev , "David S. Ahern" , KVM list , qemu-devel@nongnu.org To: Mark Marshall , Stefano Stabellini Return-path: Received: from mx2.redhat.com ([66.187.237.31]:34152 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751278AbZBDIKh (ORCPT ); Wed, 4 Feb 2009 03:10:37 -0500 Content-Disposition: inline In-Reply-To: <4988032A.5010905@ntlworld.com> Sender: kvm-owner@vger.kernel.org List-ID: On Tue, Feb 03, 2009 at 08:41:14AM +0000, Mark Marshall wrote: > Michael Tokarev wrote: >> David S. Ahern wrote: >>> Michael Tokarev wrote: >>>> David S. Ahern wrote: >>>>> I don't recall seeing a followup to this post. >>>>> >>>>> To put Michael's second suggestion into patch form, the following fixes >>>>> the problem for me: >>>>> >>>>> --- kvm-81.orig/qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 >>>>> +++ kvm-81/qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 >>>>> @@ -1078,20 +1078,21 @@ >>>>> if (sarg | TIOCM_DTR) >>>>> *targ |= CHR_TIOCM_DTR; >>>>> if (sarg | TIOCM_RTS) >>>>> *targ |= CHR_TIOCM_RTS; >>>>> } >>>>> break; >>>>> case CHR_IOCTL_SERIAL_SET_TIOCM: >>>>> { >>>>> int sarg = *(int *)arg; >>>>> int targ = 0; >>>>> + ioctl(s->fd_in, TIOCMGET, &targ); >>>> here, one more operation is necessary: >>>> targ &= ~(TIOCM_DTR|TIOCM_RTS); >>> Interesting. that change was not needed to fix my problem. >> >> It just means you (or, rather, your guests) never really needed to >> DROP those signal lines, only to raise them. >> >>>>> if (sarg | CHR_TIOCM_DTR) >>>>> targ |= TIOCM_DTR; >>>>> if (sarg | CHR_TIOCM_RTS) >>>>> targ |= TIOCM_RTS; > Is this code really correct. If it is is there a comment somewhere > describing why it's correct? I would have expected the lines above to > be: > > if (sarg & CHR_TIOCM_DTR) > targ |= TIOCM_DTR; > if (sarg & CHR_TIOCM_RTS) > targ |= TIOCM_RTS; > > (Just an observation as this went past). Right. I don't understand the point of converting to an "internal" representation of TIOCM control bits. CHR_IOCTL_SERIAL_SET_TIOCM is clearly broken as mentioned. It should at least preserve the bits it does not control. diff --git a/qemu/qemu-char.c b/qemu/qemu-char.c index ac431c7..66971e1 100644 --- a/qemu/qemu-char.c +++ b/qemu/qemu-char.c @@ -1063,33 +1063,12 @@ static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg) break; case CHR_IOCTL_SERIAL_GET_TIOCM: { - int sarg = 0; - int *targ = (int *)arg; - ioctl(s->fd_in, TIOCMGET, &sarg); - *targ = 0; - if (sarg | TIOCM_CTS) - *targ |= CHR_TIOCM_CTS; - if (sarg | TIOCM_CAR) - *targ |= CHR_TIOCM_CAR; - if (sarg | TIOCM_DSR) - *targ |= CHR_TIOCM_DSR; - if (sarg | TIOCM_RI) - *targ |= CHR_TIOCM_RI; - if (sarg | TIOCM_DTR) - *targ |= CHR_TIOCM_DTR; - if (sarg | TIOCM_RTS) - *targ |= CHR_TIOCM_RTS; + ioctl(s->fd_in, TIOCMGET, arg); } break; case CHR_IOCTL_SERIAL_SET_TIOCM: { - int sarg = *(int *)arg; - int targ = 0; - if (sarg | CHR_TIOCM_DTR) - targ |= TIOCM_DTR; - if (sarg | CHR_TIOCM_RTS) - targ |= TIOCM_RTS; - ioctl(s->fd_in, TIOCMSET, &targ); + ioctl(s->fd_in, TIOCMSET, arg); } break; default: