Re: [PATCH 2/6] e1000: use pcix_set_mmrbc

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Stephen Hemminger wrote:
On Fri, 08 Dec 2006 13:45:05 -0800
Roland Dreier <[email protected]> wrote:

 > -        if (hw->bus_type == e1000_bus_type_pcix) {
 > -            e1000_read_pci_cfg(hw, PCIX_COMMAND_REGISTER, &pcix_cmd_word);
 > -            e1000_read_pci_cfg(hw, PCIX_STATUS_REGISTER_HI,
 > -                &pcix_stat_hi_word);
 > -            cmd_mmrbc = (pcix_cmd_word & PCIX_COMMAND_MMRBC_MASK) >>
 > -                PCIX_COMMAND_MMRBC_SHIFT;
 > -            stat_mmrbc = (pcix_stat_hi_word & PCIX_STATUS_HI_MMRBC_MASK) >>
 > -                PCIX_STATUS_HI_MMRBC_SHIFT;
 > -            if (stat_mmrbc == PCIX_STATUS_HI_MMRBC_4K)
 > -                stat_mmrbc = PCIX_STATUS_HI_MMRBC_2K;
 > -            if (cmd_mmrbc > stat_mmrbc) {
 > -                pcix_cmd_word &= ~PCIX_COMMAND_MMRBC_MASK;
 > -                pcix_cmd_word |= stat_mmrbc << PCIX_COMMAND_MMRBC_SHIFT;
 > -                e1000_write_pci_cfg(hw, PCIX_COMMAND_REGISTER,
 > -                    &pcix_cmd_word);
 > -            }
 > -        }
 > +        if (hw->bus_type == e1000_bus_type_pcix)
 > +		e1000_pcix_set_mmrbc(hw, 2048);

This changes the behavior of the driver.  The existing driver only
sets MMRBC if it's bigger than min(2048, value in the status register).
You're setting MMRBC to 2048 even if it starts out at a smaller value.

 - R.

Hmm.. looks like all that code should really be moved off to PCI bus
quirk/setup.  None of it is E1000 specific.  Something like this (untested):

This is not true, and I have to NAK the original patch. Part of the code Stephan is removing fixes a BUG in one of our *e1000 parts* that has the wrong size.

It would be nice to fix generix pci-x issues qith quirks for platforms but the adjustment needs to stay for this specific e1000 case.

Perhaps we can accomodate that specific case so that it is apparent from our code, as is not the case right now.

Auke

PS Thanks to Jeb for fishing this out ;)



---
 drivers/pci/quirks.c |   30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

--- pci-x.orig/drivers/pci/quirks.c
+++ pci-x/drivers/pci/quirks.c
@@ -1719,6 +1719,36 @@ DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_NV
 #endif /* CONFIG_PCI_MSI */
EXPORT_SYMBOL(pcie_mch_quirk);
+
+/* Check that BIOS has not set PCI-X MMRBC to value bigger than board allows */
+static void __devinit quirk_pcix_mmrbc(struct pci_dev *dev)
+{
+	int cap;
+	u32 stat;
+	u16 cmd, m, c;
+
+	cap = pci_find_capability(dev, PCI_CAP_ID_PCIX);
+	if (cap <= 0)
+		return;
+
+	if (pci_read_config_dword(dev, cap + PCI_X_STATUS, &stat) ||
+	    pci_read_config_word(dev, cap + PCI_X_CMD, &cmd))
+		return;
+
+	m = (stat & PCI_X_STATUS_MAX_READ) >> 21;	/* max possible MRRBC*/
+	c = (cmd & PCI_X_CMD_MAX_READ) >> 2;		/* current MMRBC */
+	if (c > m) {
+		printk(KERN_INFO "PCIX: %s resetting MMRBC to %d\n",
+		       pci_name(dev), 512 << m);
+
+		cmd &= ~PCI_X_CMD_MAX_READ;
+		cmd |= m << 2;
+		pci_write_config_dword(dev, cap + PCI_X_CMD, cmd);
+	}
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_ANY_ID, PCI_ANY_ID, quirk_pcix_mmrbc);
+
+
 #ifdef CONFIG_HOTPLUG
 EXPORT_SYMBOL(pci_fixup_device);
 #endif


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux