>-----Original Message----- >From: Jesse Barnes [mailto:jbarnes@xxxxxxxxxxxxxxxx] >Sent: Tuesday, December 09, 2008 2:52 PM >To: Pan, Jacob jun >Cc: linux-pci@xxxxxxxxxxxxxxx >Subject: Re: [PATCH] Avoid calling early PCI mmconfig init if pci=noearly is >given in cmdline > >On Monday, December 08, 2008 9:59 am Pan, Jacob jun wrote: >> From 3bcb87f05ee8f492d83a5bb1fc43c639c1abd784 Mon Sep 17 00:00:00 2001 >> From: Jacob Pan <jacob.jun.pan@xxxxxxxxx> >> Date: Mon, 8 Dec 2008 09:44:16 -0800 >> Subject: [PATCH] Avoid calling early PCI mmconfig init if pci=noearly is >> given in cmdline >> >> >> Signed-off-by: Jacob Pan <jacob.jun.pan@xxxxxxxxx> > >Jacob, can you include some info about why we need this change and why it's >safe? [[JPAN]] the problem is that without this change, pci_mmcfg_early_init() would call type1 config reads, it causes trouble for the chip I am working on. There would be no effect for pci=noearly flag w/o this change. It would be safe in that, for both cases: -w/o pci=noearly, the execution flow would be unchanged -w/ pci=noearly, people would expect no type I early access as documented in kernel-parameters.txt. pci=noearly is probably not very common, but we don't want to regress >any machines that might be relying on it. > [[JPAN]] the fact that nobody complained so far might be just no one else uses this flag. >Jesse -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
