Anton, I understand, and I would also like to know the root cause, but we would need this logging to understand the root cause anyway. It must be some kind of timing/cache issue, given the test results. With this patch applied, if the issue shows up again, we should be able to find root cause.
Given the low importance of this issue, and the fact that it works with it applied, this patch is all that is reasonable to do about it right now, IMO. Andy -----Original Message----- From: Anton Pak [mailto:[email protected]] Sent: Thursday, July 01, 2010 8:53 AM To: [email protected] Subject: Re: [Openhpi-devel] Patch for ProcessFru/SetAddr issue Andy, Seems your patch merely adds debug logging. I am curious how it is supposed to fix the issue. Anton Pak On Thu, 01 Jul 2010 16:43:44 +0400, Andy Cress <[email protected]> wrote: > Folks, > > Even though it wasn't important enough to delay the openhpi-2.15 > release, I do have a resolution to the issue I found in testing, below. > > 5) hpiinv returned the FRU data for the baseboard (sa=0x20) instead of > the front panel board (sa=0xc0). > This was an old problem, not something new. > This should have been handled in > ipmi_mc_vendor_intel.cpp/cIpmiMcVendorIntelBmc::ProcessFru by SetAddr(). > That code executed as designed, but did not have the desired effect. > Not Important - a pre-existing issue, only on certain platforms > > The code seemed to be correct for SetAddr() to change the slave_addr and > for it to be used in reading the FRU, and in fact, when I added code to > log what it was really doing, it did work. With this patch, it works, > with or without logging. Without this patch, it doesn't properly Read > the FRU from sa=0xc0. > Without this logic it would not be possible to root cause the problem > anyway, so I propose that this be added as a patch. > > Andy > > --- plugins/ipmidirect/ipmi_con_smi.cpp-old 2010-06-30 > 14:26:36.000000000 -0400 > +++ plugins/ipmidirect/ipmi_con_smi.cpp 2010-07-01 08:05:35.000000000 > -0400 > @@ -327,6 +327,13 @@ > > req.msgid = r->m_seq; > > + if (send_addr.m_slave_addr != 0x20) { > + stdlog << "SMI SendCmd: sa=" << send_addr.m_slave_addr << > + " chan=" << send_addr.m_channel << > + " type=" << send_addr.m_type << > + " netfn=" << req.msg.netfn << > + " cmd=" << req.msg.cmd << ".\n"; > + } > int rv = ioctl( m_fd, IPMICTL_SEND_COMMAND, &req ); > > if ( rv ) > > > ------------------------------------------------------------------------ ------ > This SF.net email is sponsored by Sprint > What will you do first with EVO, the first 4G phone? > Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first > _______________________________________________ > Openhpi-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/openhpi-devel ------------------------------------------------------------------------ ------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Openhpi-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openhpi-devel ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Openhpi-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openhpi-devel
