On Wed, Mar 26, 2008 at 11:33:24AM -0700, Linus Torvalds wrote:Could be but without the sizing removal I doubt that the fix I was going to propose (see below) would be the same. What I had found is that the removal of the transparent bridge sizing was leaving the resource record for at least region 7 (IO) of a hot-added transparent bridge on the docking station in a state that was not palatable with later executed code in pdev_sort_resource(). Even though the restoration of the transparent bridge sizing corrects the problem, pdev_sort_resource() could probably use some bulletproofing. I will take a look at this. Thanks, Gary -- Gary Hade System x Enablement IBM Linux Technology Center 503-578-4503 IBM T/L: 775-4503 garyhade@us.ibm.com http://www.ibm.com/linux/ltc --- linux-2.6.25-rc6/drivers/pci/probe.c.orig 2008-03-20 12:09:14.000000000 -0700 +++ linux-2.6.25-rc6/drivers/pci/probe.c 2008-03-20 12:11:29.000000000 -0700 @@ -328,6 +328,8 @@ void __devinit pci_read_bridge_bases(str if (!res->end) res->end = limit + 0xfff; } + if (!res->start && dev->transparent) + res->flags = 0; res = child->resource[1]; pci_read_config_word(dev, PCI_MEMORY_BASE, &mem_base_lo); @@ -339,6 +341,8 @@ void __devinit pci_read_bridge_bases(str res->start = base; res->end = limit + 0xfffff; } + if (!res->start && dev->transparent) + res->flags = 0; res = child->resource[2]; pci_read_config_word(dev, PCI_PREF_MEMORY_BASE, &mem_base_lo); @@ -373,6 +377,8 @@ void __devinit pci_read_bridge_bases(str res->start = base; res->end = limit + 0xfffff; } + if (!res->start && dev->transparent) + res->flags = 0; } static struct pci_bus * pci_alloc_bus(void) --
| Andy Whitcroft | clam |
| Tarkan Erimer | Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3 |
| Greg Kroah-Hartman | [PATCH 001/196] Chinese: Add the known_regression URI to the HOWTO |
| David Miller | Re: Slow DOWN, please!!! |
git: | |
| Arjan van de Ven | Re: [GIT]: Networking |
| Lennert Buytenhek | [PATCH 08/39] mv643xx_eth: nuke port status register bit defines |
| Jarek Poplawski | Re: HTB accuracy for high speed |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
