Sort generic PCI fixups after specific ones]
Subsystems (3)
PATCH 1/1 Sort generic PCI fixups after specific ones]
Greg asked me to resend this patch and cc l-k, so here it is.
----- Forwarded message from Ivan Kokshaysky <[email protected]> -----
Date: Sat, 2 Oct 2004 01:48:17 +0400
From: Ivan Kokshaysky <[email protected]>
To: Matthew Wilcox <[email protected]>
Cc: Greg KH <[email protected]>, Linus Torvalds <[email protected]>,
Andrew Morton <[email protected]>, [email protected],
[email protected], [email protected]
Subject: Re: [PATCH] Sort generic PCI fixups after specific ones
User-Agent: Mutt/1.2.5i
In-Reply-To: <[email protected]>; from [email protected] on Thu, Sep 30, 2004 at 06:41:55PM +0100
X-Spam-Score: 0.0 (/)
Quoted 3 lines · Matthew Wilcox
Allow prioritising PCI fixups. "How it works" is covered in the comment in
pci.h. The patch tosuperio.cmay well only apply with fuzz to the current Linux tree; I include it only to show an example.
No, you missed my point.
What we need is yet another PCI fixup *pass*, not prioritizing fixups inside *one* pass - see appended patch (compile tested only). Speaking of IDE: generally, you cannot switch PCI IDE controller from "compatible" to native mode in the "fixup header" pass, as at this point PCI BARs are *already* probed, and in compatible mode a lot of IDE controllers have some bogus (often read-only) values in the BARs 0-3. That's why quirk_ide_bases() fixup exists in the first place. So that parisc superio fix works by mere luck - probably because IDE BARs 0-3 are readable/writable even in legacy mode on this particular controller.
Ivan.
include/asm-generic/vmlinux.lds.h +3 -0
@@ -18,6 +18,9 @@ \/* PCI quirks */ \ .pci_fixup : AT(ADDR(.pci_fixup) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start_pci_fixups_early) = .; \ + *(.pci_fixup_early) \ + VMLINUX_SYMBOL(__end_pci_fixups_early) = .; \ VMLINUX_SYMBOL(__start_pci_fixups_header) = .; \ *(.pci_fixup_header) \ VMLINUX_SYMBOL(__end_pci_fixups_header) = .; \
include/linux/pci.h +18 -9
@@ -1006,20 +1006,29 @@ struct pci_fixup { };enum pci_fixup_pass { - pci_fixup_header,/* Called immediately after reading configuration header */ + pci_fixup_early,/* Called immediately after reading +device/vendor IDs and class code */ + pci_fixup_header,/* Called after reading the entire +configuration header (including BARs) */ pci_fixup_final,/* Final phase of device fixups */ };/* Anonymous variables would be nice... */ -#define DECLARE_PCI_FIXUP_HEADER (vendor, device, hook) \ - static struct pci_fixup __pci_fixup_##vendor##device##hook __attribute_used__ \ - __attribute__((__section__(".pci_fixup_header"))) = { \ - vendor, device, hook }; +#define DECLARE_PCI_FIXUP_SECTION (section, name, vendor, device, hook) \ + static struct pci_fixup __pci_fixup_##name __attribute_used__ \ + __attribute__((__section__( #section ))) = { vendor, device, hook }; -#define DECLARE_PCI_FIXUP_FINAL (vendor, device, hook) \ - static struct pci_fixup __pci_fixup_##vendor##device##hook __attribute_used__ \ - __attribute__((__section__(".pci_fixup_final"))) = { \ - vendor, device, hook }; +#define DECLARE_PCI_FIXUP_EARLY (vendor, device, hook) \ + DECLARE_PCI_FIXUP_SECTION(.pci_fixup_early, \ + vendor##device##hook, vendor, device, hook) + +#define DECLARE_PCI_FIXUP_HEADER (vendor, device, hook) \ + DECLARE_PCI_FIXUP_SECTION(.pci_fixup_header, \ + vendor##device##hook, vendor, device, hook) + +#define DECLARE_PCI_FIXUP_FINAL (vendor, device, hook) \ + DECLARE_PCI_FIXUP_SECTION(.pci_fixup_final, \ + vendor##device##hook, vendor, device, hook)void pci_fixup_device (enum pci_fixup_pass pass,struct pci_dev *dev);
drivers/pci/quirks.c +7 -0
@@ -999,6 +999,8 @@ static void pci_do_fixups(struct pci_dev } } +extern struct pci_fixup __start_pci_fixups_early[]; +extern struct pci_fixup __end_pci_fixups_early[];extern struct pci_fixup __start_pci_fixups_header[];extern struct pci_fixup __end_pci_fixups_header[];extern struct pci_fixup __start_pci_fixups_final[];
@@ -1009,6 +1011,11 @@ void pci_fixup_device(enum pci_fixup_passtruct pci_fixup *start, *end;switch (pass) { +case pci_fixup_early: + start = __start_pci_fixups_early; + end = __end_pci_fixups_early; +break ; +case pci_fixup_header: start = __start_pci_fixups_header; end = __end_pci_fixups_header;
drivers/pci/probe.c +3 -0
@@ -475,6 +475,9 @@ static int pci_setup_device(struct pci_d/* "Unknown power state" */ dev->current_state =4 ; +/* Early fixups, before probing the BARs */ +pci_fixup_device (pci_fixup_early, dev); +switch (dev->hdr_type) {/* header type */ case PCI_HEADER_TYPE_NORMAL :/* standard header */ if (class ==PCI_CLASS_BRIDGE_PCI )
drivers/parisc/superio.c +1 -1
@@ -484,7 +484,7 @@ void superio_fixup_pci(struct pci_dev *ppci_read_config_byte (pdev,PCI_CLASS_PROG , &prog);printk ("PCI: Enabled native mode for NS87415 (pif=0x%x)\n" , prog); } -DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_NS ,PCI_DEVICE_ID_NS_87415 , superio_fixup_pci); +DECLARE_PCI_FIXUP_EARLY (PCI_VENDOR_ID_NS ,PCI_DEVICE_ID_NS_87415 , superio_fixup_pci);/* Because of a defect in Super I/O, all reads of the PCI DMA status * registers, IDE status register and the IDE select register need to be
----- End forwarded message -----
Signature · 6 lines
"Next the statesmen will invent cheap lies, putting the blame upon the nation that is attacked, and every man will be glad of those conscience-soothing falsities, and will diligently study them, and refuse to examine any refutations of them; and thus he will by and by convince himself that the war is just, and will thank God for the better sleep he enjoys after this process of grotesque self-deception." -- Mark Twain
Quoted 2 lines · Matthew Wilcox
Greg asked me to resend this patch and cc l-k, so here it is.
Does this make ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.9-rc3/2.6.9-rc3-mm2/broken-out/sort-generic-pci-fixups-after-specific-ones.patch obsolete, or what?
Quoted 7 lines · Andrew Morton
Quoted 2 lines · Matthew Wilcox
Greg asked me to resend this patch and cc l-k, so here it is.
Does this make ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.9-rc3/2.6.9-rc3-mm2/broken-out/sort-generic-pci-fixups-after-specific-ones.patch obsolete, or what?
Yes. Ivan thinks there are cases that sort-generic-pci-fixups-after-specific-ones.patch won't fix, and I don't care enough to argue. His patch fixes my problem, so let's go with it instead.
Signature · 6 lines
"Next the statesmen will invent cheap lies, putting the blame upon the nation that is attacked, and every man will be glad of those conscience-soothing falsities, and will diligently study them, and refuse to examine any refutations of them; and thus he will by and by convince himself that the war is just, and will thank God for the better sleep he enjoys after this process of grotesque self-deception." -- Mark Twain