linux-kernel - Re: [PATCH v2] xen-pciback: optionally allow interrupt enable flag writes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 19 Dec 2019 12:20:24 +0100
From:   Jan Beulich <jbeulich@...e.com>
To:     Marek Marczykowski-Górecki 
        <marmarek@...isiblethingslab.com>
Cc:     xen-devel@...ts.xenproject.org,
        Roger Pau Monné <roger.pau@...rix.com>,
        YueHaibing <yuehaibing@...wei.com>,
        Simon Gaiser <simon@...isiblethingslab.com>,
        Stefano Stabellini <sstabellini@...nel.org>,
        Boris Ostrovsky <boris.ostrovsky@...cle.com>,
        Juergen Gross <jgross@...e.com>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] xen-pciback: optionally allow interrupt enable flag
 writes

On 19.12.2019 04:49, Marek Marczykowski-Górecki  wrote:
> +enum interrupt_type xen_pcibk_get_interrupt_type(struct pci_dev *dev)
> +{
> +	int err;
> +	u16 val;
> +
> +	err = pci_read_config_word(dev, PCI_COMMAND, &val);
> +	if (err)
> +		return INTERRUPT_TYPE_ERR;
> +	if (!(val & PCI_COMMAND_INTX_DISABLE))
> +		return INTERRUPT_TYPE_INTX;
> +
> +	/* Do not trust dev->msi(x)_enabled here, as enabling could be done
> +	 * bypassing the pci_*msi* functions, by the qemu.
> +	 */

Judging from this comment, how can you assume only one of the
three variants is actually enabled? It's against the spec, yes,
but it's not at all impossible afaict. I think you want the
return value here to be
- negative errno values (no need to discard the actual error
  codes) or
- a non-negative bitmap indicating which of the interrupt types
  is/are currently enabled.
That way ...

> +static int msi_msix_flags_write(struct pci_dev *dev, int offset, u16 new_value,
> +				void *data)
> +{
> +	int err;
> +	u16 old_value;
> +	const struct msi_msix_field_config *field_config = data;
> +	const struct xen_pcibk_dev_data *dev_data = pci_get_drvdata(dev);
> +
> +	if (xen_pcibk_permissive || dev_data->permissive)
> +		goto write;
> +
> +	err = pci_read_config_word(dev, offset, &old_value);
> +	if (err)
> +		return err;
> +
> +	if (new_value == old_value)
> +		return 0;
> +
> +	if (!dev_data->allow_interrupt_control ||
> +	    (new_value ^ old_value) & ~field_config->enable_bit)
> +		return PCIBIOS_SET_FAILED;
> +
> +	if (new_value & field_config->enable_bit) {
> +		/* don't allow enabling together with other interrupt types */
> +		const enum interrupt_type int_type = xen_pcibk_get_interrupt_type(dev);
> +		if (int_type == INTERRUPT_TYPE_NONE ||
> +		    int_type == field_config->int_type)

... equality comparisons like this one will actually become safe.

Jan