> -----Original Message----- > From: linux-omap-owner@xxxxxxxxxxxxxxx > [mailto:linux-omap-owner@xxxxxxxxxxxxxxx] On Behalf Of C.A, > Subramaniam > Sent: Thursday, June 24, 2010 4:36 PM > To: Ohad Ben-Cohen; Kanigeri, Hari > Cc: linux-omap@xxxxxxxxxxxxxxx; Hiroshi Doyu; Ramirez Luna, Omar > Subject: RE: [PATCH 3/4] omap mailbox: remove mbox_configured scheme > > > > -----Original Message----- > > From: linux-omap-owner@xxxxxxxxxxxxxxx > > [mailto:linux-omap-owner@xxxxxxxxxxxxxxx] On Behalf Of Ohad > Ben-Cohen > > Sent: Thursday, June 24, 2010 3:23 PM > > To: Kanigeri, Hari > > Cc: linux-omap@xxxxxxxxxxxxxxx; Hiroshi Doyu; Ramirez Luna, Omar > > Subject: Re: [PATCH 3/4] omap mailbox: remove mbox_configured scheme > > > > On Thu, Jun 24, 2010 at 6:10 PM, Kanigeri, Hari > <h-kanigeri2@xxxxxx> > > wrote: > > > Ohad, > > > > > >> mbox_configured is global and therefore does not support > > concurrent > > >> usage of multiple mailbox instances. > > > > > > -- Why do you say that it doesn't support concurrent usage > > of multiple mailbox instances ? If you take example of > OMAP4, we have > > 2 mailbox instances, one talking to DSP and other to Ducati > and they > > should be supported concurrently. > > > > > > Interesting. was this actually tested and found working ? > > > I think by supporting multiple instances, what Hari meant was > that, for DSP and Ducati 2 different struct omap_mbox * are > returned. They still can maintain their own callback function > (http://dev.omapzoom.org/?p=tisyslink/kernel-syslink.git;a=blo > b;f=arch/arm/mach-omap2/mailbox.c;h=a23c7edf1e84ab4fa51e56d0c2 > daf2922084751a;hb=438f7a6a3cced478eb121426201206f6205fbbdc#l32 > 7 for Ducati and > > http://dev.omapzoom.org/?p=tisyslink/kernel-syslink.git;a=blob > ;f=arch/arm/mach-omap2/mailbox.c;h=a23c7edf1e84ab4fa51e56d0c2d > af2922084751a;hb=438f7a6a3cced478eb121426201206f6205fbbdc#l334 > for DSP). > > The "mbox_configured" was a means to reference count for one > instance of the struct omap_mbox * . Sorry I that that back. Even though we maintain 2 strucutres, only the one that calls the request_irq() is honoured. Hence yes only one callback will be serviced (Thank to Ramesh for pointing that out). Chaining of callbacks is a good way to allow multiple clients to be notified. > > > AFAICT (I don't have OMAP4, this is just by looking at the > > code) this doesn't work: > > mbox_configured will prevent omap2_mbox_startup from being invoked > > more than once. This means that omap2_mbox_enable_irq(mbox, IRQ_RX) > > will only be called once too, which seems like a problem if > you want > > to start receiving RX interrupts from both mbox instances > at the same > > time. > > > > To fix that I guess we should decouple the RX interrupt > enabling from > > the mbox startup function. I can write something, but I'd need your > > help to test it on OMAP4 :) > > > > > > > > If you remove the mbox_configured variable, then the > mailbox module > > > would shut down once the first instance calls the > > omap_mbox_put function. > > > > Also interesting point. > > > > So you use mbox_configures as a reference counter to allow > concurrent > > multiple senders to the same mbox instance. > > > > I will update the previous patch (that introduced the atomic_t > > reference counter) to reflect this use case, thanks! > > > > > > > > Thank you, > > > Best regards, > > > Hari > > > > > -- > > To unsubscribe from this list: send the line "unsubscribe > linux-omap" > > in the body of a message to majordomo@xxxxxxxxxxxxxxx More > majordomo > > info at http://vger.kernel.org/majordomo-info.html > > > > > Thank you and Regards > Subbu-- > To unsubscribe from this list: send the line "unsubscribe > linux-omap" in the body of a message to > majordomo@xxxxxxxxxxxxxxx More majordomo info at > http://vger.kernel.org/majordomo-info.html > Thank you and Regards Subbu-- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
