openwrt/target/linux/bcm47xx/patches-5.4/900-ssb-reject-PCI-writes-setting-CardBus-bridge-resourc.patch
Rafał Miłecki d369d92fc7 bcm47xx: add support for kernel 5.4
Ethernet, switch, LEDs, buttons, USB, sysupgrade & LuCI were
successfully tested on BCM4706.

WARNING: Hack for BCM4710 adding BCM4710_PROTECTED_FILL_TLB() to the
local_r4k_flush_cache_sigtramp() could not be ported. That function has
been dropped in:
commit adcc81f148d7 ("MIPS: math-emu: Write-protect delay slot emulation pages")
commit 3315b6b336c8 ("MIPS: Delete unused flush_cache_sigtramp()")
it's unsure if that chipset will still work reliably.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2020-03-10 15:50:55 +01:00

31 lines
1.1 KiB
Diff

From 5c81397a0147ea59c778d1de14ef54e2268221f6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <zajec5@gmail.com>
Date: Wed, 8 Apr 2015 06:58:11 +0200
Subject: [PATCH] ssb: reject PCI writes setting CardBus bridge resources
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
If SoC has a CardBus we can set resources of device at slot 1 only. It's
impossigle to set bridge resources as it simply overwrites device 1
configuration and usually results in Data bus error-s.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
---
drivers/ssb/driver_pcicore.c | 4 ++++
1 file changed, 4 insertions(+)
--- a/drivers/ssb/driver_pcicore.c
+++ b/drivers/ssb/driver_pcicore.c
@@ -164,6 +164,10 @@ static int ssb_extpci_write_config(struc
WARN_ON(!pc->hostmode);
if (unlikely(len != 1 && len != 2 && len != 4))
goto out;
+ /* CardBus SoCs allow configuring dev 1 resources only */
+ if (extpci_core->cardbusmode && dev != 1 &&
+ off >= PCI_BASE_ADDRESS_0 && off <= PCI_BASE_ADDRESS_5)
+ goto out;
addr = get_cfgspace_addr(pc, bus, dev, func, off);
if (unlikely(!addr))
goto out;