[PATCH net v2] e1000e: Fix out-of-bounds MMIO access by validating BAR0 size

[PATCH net v2] e1000e: Fix out-of-bounds MMIO access by validating BAR0 size

[Pu Lehui]

From: Pu Lehui <pulehui@huawei.com>

Syzkaller reported a kernel panic caused by an out-of-bounds MMIO
access in the e1000e driver.

[   82.868719][  T404] e1000e 0000:00:02.0: The NVM Checksum Is Not Valid
[   82.872328][  T404] Unable to handle kernel paging request at virtual address ffff80008894e090
[   83.085218][  T404] CPU: 2 UID: 0 PID: 404 Comm: bash Not tainted 7.2.0-rc2-g3f1f75536668 #1 PREEMPTLAZY
[   83.129013][  T404] pc : e1000_get_cfg_done_82571+0x70/0x158
[   83.140092][  T404] lr : e1000_get_cfg_done_82571+0x68/0x158
[   83.151196][  T404] sp : ffff80008ac37410
[   83.158922][  T404] x29: ffff80008ac37410 x28: ffff0000cd6a11b8 x27: ffff0000c58190d0
[   83.173919][  T404] x26: ffff0000cd6a11b8 x25: ffff0000cd6a0bc0 x24: ffff0000cd6a0000
[   83.189417][  T404] x23: 0000000000001010 x22: ffff0000cd6a11c0 x21: ffff0000cd6a11b8
[   83.205195][  T404] x20: 0000000000000064 x19: ffff80008894e090 x18: 0000000000000000
[   83.220545][  T404] x17: ffff800081c1a3f4 x16: ffff800081c19c10 x15: ffff800081e86510
[   83.235764][  T404] x14: 0000000000000001 x13: 0000000000000001 x12: ffff60001bc8a8b3
[   83.251301][  T404] x11: 1fffe0001bc8a8b2 x10: ffff60001bc8a8b2 x9 : ffff800081eae25c
[   83.266705][  T404] x8 : 00009fffe437574e x7 : ffff0000de454593 x6 : 0000000000000001
[   83.281919][  T404] x5 : ffff0000cf2b9640 x4 : 0000000000000000 x3 : dfff800000000000
[   83.297317][  T404] x2 : 0000000000000007 x1 : ffff0000cd6a11c0 x0 : 0000000000000000
[   83.312601][  T404] Call trace:
[   83.318662][  T404]  e1000_get_cfg_done_82571+0x70/0x158 (P)
[   83.329748][  T404]  e1000e_phy_hw_reset_generic+0x17c/0x1a8
[   83.341541][  T404]  e1000_probe+0xbd8/0x1988
[   83.350334][  T404]  local_pci_probe+0x84/0x130

Repetition steps:
1. Find PCI device which BAR0 size <= 4K. If it's:
   Device Addr: 0000:00:02.0   BAR0 SIZE: 4K
   Vendor/Device ID: 0x1af4 0x1004
2. Unbind the above PCI device
   echo '0000:00:02.0' > /sys/bus/pci/devices/0000:00:02.0/driver/unbind
3. Set the above device to e1000e new_id
   echo '1af4 1004' > /sys/bus/pci/drivers/e1000e/new_id

During e1000_probe(), the driver maps the device's BAR0 memory region.
If the device has a 4K BAR0, ioremap() maps only 4K of space. Later in
the probe process, when the NVM checksum validation fails, the driver
attempts to perform a hardware reset and falls back to the err_eeprom
cleanup path.

This cleanup path will trigger an OOB access kernel panic:
e1000_phy_hw_reset
  e1000e_phy_hw_reset_generic
    e1000_get_cfg_done_82571
      er32(EEMNGCTL)
        readl(hw->hw_addr + EEMNGCTL); <-- EEMNGCTL(0x1010) > 4K, OOB access

Fix this by ensuring the MMIO length (pci_resource_len(pdev, 0)) is at
least 64K (0x10000) before proceeding with ioremap(). The 64K minimum
safely covers the maximum register offset accessed by the e1000e driver.

Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)")
Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
v2:
- Update comment about E1000_MMIO_LEN_MIN. (Breno)

 drivers/net/ethernet/intel/e1000e/netdev.c | 5 +++++
 drivers/net/ethernet/intel/e1000e/regs.h   | 3 +++
 2 files changed, 8 insertions(+)

diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index 844f31ab37ad..46e58e926f7b 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -7450,6 +7450,11 @@ static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	mmio_len = pci_resource_len(pdev, 0);
 
 	err = -EIO;
+	if (mmio_len < E1000_MMIO_LEN_MIN) {
+		dev_err(&pdev->dev, "MMIO len is too small\n");
+		goto err_ioremap;
+	}
+
 	adapter->hw.hw_addr = ioremap(mmio_start, mmio_len);
 	if (!adapter->hw.hw_addr)
 		goto err_ioremap;
diff --git a/drivers/net/ethernet/intel/e1000e/regs.h b/drivers/net/ethernet/intel/e1000e/regs.h
index 6c0cd8cab3ef..dbabd35f5a12 100644
--- a/drivers/net/ethernet/intel/e1000e/regs.h
+++ b/drivers/net/ethernet/intel/e1000e/regs.h
@@ -242,4 +242,7 @@
 /* PHY registers */
 #define I82579_DFT_CTRL	PHY_REG(769, 20)
 
+/* Smallest BAR0 that covers every register the driver accesses */
+#define E1000_MMIO_LEN_MIN	0x10000
+
 #endif
-- 
2.34.1

Re: [PATCH net v2] e1000e: Fix out-of-bounds MMIO access by validating BAR0 size

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The driver unconditionally maps BAR0 with `ioremap()`, without verifying that BAR0 is an `IORESOURCE_MEM` region. A spoofed PCI device could expose BAR0 as an I/O region, which could bypass the new length check on some architectures and lead to `ioremap()` being called on an I/O port address.
--

--- Patch [1]: [PATCH net v2] e1000e: Fix out-of-bounds MMIO access by validating BAR0 size ---
commit 939eee77005bb32587689d8e06aa2cec8c38f7e9
Author: Pu Lehui <pulehui@huawei.com>

e1000e: Fix out-of-bounds MMIO access by validating BAR0 size

This commit fixes an out-of-bounds MMIO access in the e1000e driver
by ensuring the length of BAR0 is at least 64K before performing ioremap.

> diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
> index 844f31ab37ad4..46e58e926f7b4 100644
> --- a/drivers/net/ethernet/intel/e1000e/netdev.c
> +++ b/drivers/net/ethernet/intel/e1000e/netdev.c

[ ... ]

> @@ -7450,6 +7450,11 @@ static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>  	mmio_len = pci_resource_len(pdev, 0);
>  
>  	err = -EIO;
> +	if (mmio_len < E1000_MMIO_LEN_MIN) {
> +		dev_err(&pdev->dev, "MMIO len is too small\n");
> +		goto err_ioremap;
> +	}
> +
>  	adapter->hw.hw_addr = ioremap(mmio_start, mmio_len);

[Severity: High]
This isn't a bug introduced by this patch, but is it possible for ioremap() 
to be called on an I/O port region here?

Earlier in e1000_probe(), pci_select_bars() filters out non-memory BARs. If 
a device is bound via sysfs and exposes BAR0 as an IORESOURCE_IO region 
instead of IORESOURCE_MEM, the driver fails to request the region but still 
unconditionally passes mmio_start to ioremap() at this location.

Could this cause a kernel panic on architectures where I/O space is distinct 
from memory space when the returned address is subsequently dereferenced?

>  	if (!adapter->hw.hw_addr)
>  		goto err_ioremap;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133049.1289167-1-pulehui@huaweicloud.com?part=1

Re: [PATCH net v2] e1000e: Fix out-of-bounds MMIO access by validating BAR0 size

[Andrew Lunn]

> --- a/drivers/net/ethernet/intel/e1000e/regs.h
> +++ b/drivers/net/ethernet/intel/e1000e/regs.h
> @@ -242,4 +242,7 @@
>  /* PHY registers */
>  #define I82579_DFT_CTRL	PHY_REG(769, 20)
>  
> +/* Smallest BAR0 that covers every register the driver accesses */
> +#define E1000_MMIO_LEN_MIN	0x10000

I commented on the previous version. Is I82579_DFT_CTRL the highest
register? Can E1000_MMIO_LEN_MIN defines based on that?

	  Andrew