Re: ioremap warning fix for jsm driver.
From: Lennart Sorensen <hidden>
Date: 2011-06-01 20:29:29
Also in:
lkml
On Wed, Jun 01, 2011 at 02:38:41PM -0400, Lennart Sorensen wrote:
I saw a warning about ioremap from the jsm driver on a system which looked like this: resource map sanity check conflict: 0xe0200800 0xe02017ff 0xe0200800 0xe0200fff 0000:01:08.0 Turns out the warning is valid. The jsm driver has been asking to ioremap 0x1000 forever, but in fact only 8 port chips have 0x1000 bytes of memory. 4 port chips have 0x800 and 2 port chips have 0x400 according to the data sheet. It makes more sense to map the size of the region rather than a hard coded value. If you happen to have the regiong legitimately mapped to a base addres that is not 4K aligned, ioremap complains otherwise.
I can't believe I didn't spell check that. :( Should have been: Turns out the warning is valid. The jsm driver has been asking to ioremap 0x1000 forever, but in fact only 8 port chips have 0x1000 bytes of memory. 4 port chips have 0x800 and 2 port chips have 0x400 according to the data sheet. It makes more sense to map the size of the region rather than a hard coded value. If you happen to have the region legitimately mapped to a base address that is not 4K aligned ioremap complains, and otherwise memory not belonging to the device is reserved as far as I can tell.
quoted hunk ↗ jump to hunk
Signed-off-by: Len Sorensen <redacted>diff --git a/drivers/tty/serial/jsm/jsm_driver.c b/drivers/tty/serial/jsm/jsm_driver.c index 18f5484..96da178 100644 --- a/drivers/tty/serial/jsm/jsm_driver.c +++ b/drivers/tty/serial/jsm/jsm_driver.c@@ -125,7 +125,7 @@ static int __devinit jsm_probe_one(struct pci_dev *pdev, const struct pci_device brd->bd_uart_offset = 0x200; brd->bd_dividend = 921600; - brd->re_map_membase = ioremap(brd->membase, 0x1000); + brd->re_map_membase = ioremap(brd->membase, pci_resource_len(pdev, 0)); if (!brd->re_map_membase) { dev_err(&pdev->dev, "card has no PCI Memory resources, "-- Len Sorensen