linux-kernel - Re: [PATCH v3 3/6] PCI: endpoint: Add RC-to-EP doorbell support using platform MSI controller

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]

Message-ID: <87bjzkau33.ffs@tglx>
Date: Wed, 16 Oct 2024 18:30:40 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Frank Li <Frank.Li@....com>, Manivannan Sadhasivam
 <manivannan.sadhasivam@...aro.org>, Krzysztof Wilczyński
 <kw@...ux.com>,
 Kishon Vijay Abraham I <kishon@...nel.org>, Bjorn Helgaas
 <bhelgaas@...gle.com>, Arnd Bergmann <arnd@...db.de>, Greg Kroah-Hartman
 <gregkh@...uxfoundation.org>
Cc: linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org,
 imx@...ts.linux.dev, Niklas Cassel <cassel@...nel.org>,
 dlemoal@...nel.org, maz@...nel.org, jdmason@...zu.us, Frank Li
 <Frank.Li@....com>
Subject: Re: [PATCH v3 3/6] PCI: endpoint: Add RC-to-EP doorbell support
 using platform MSI controller

On Tue, Oct 15 2024 at 18:07, Frank Li wrote:
> +static int pci_epc_alloc_doorbell(struct pci_epc *epc, u8 func_no, u8 vfunc_no, int num_db)
> +{
> +	struct msi_desc *desc, *failed_desc;
> +	struct pci_epf *epf;
> +	struct device *dev;
> +	int i = 0;
> +	int ret;
> +
> +	if (IS_ERR_OR_NULL(epc))
> +		return -EINVAL;
> +
> +	/* Currently only support one func and one vfunc for doorbell */
> +	if (func_no || vfunc_no)
> +		return -EINVAL;
> +
> +	epf = list_first_entry_or_null(&epc->pci_epf, struct pci_epf, list);
> +	if (!epf)
> +		return -EINVAL;
> +
> +	dev = epc->dev.parent;
> +	ret = platform_device_msi_init_and_alloc_irqs(dev, num_db, pci_epc_write_msi_msg);
> +	if (ret) {
> +		dev_err(dev, "Failed to allocate MSI\n");
> +		return -ENOMEM;
> +	}
> +
> +	scoped_guard(msi_descs, dev)
> +		msi_for_each_desc(desc, dev, MSI_DESC_ALL) {

That's just wrong. Nothing in this code has to fiddle with MSI
descriptors or the descriptor lock.

        for (i = 0; i < num_db; i++) {
            virq = msi_get_virq(dev, i);

> +			ret = request_irq(desc->irq, pci_epf_doorbell_handler, 0,
> +					  kasprintf(GFP_KERNEL, "pci-epc-doorbell%d", i++), epf);
> +			if (ret) {
> +				dev_err(dev, "Failed to request doorbell\n");
> +				failed_desc = desc;
> +				goto err_request_irq;
> +			}
> +		}
> +
> +	return 0;
> +
> +err_request_irq:
> +	scoped_guard(msi_descs, dev)
> +		msi_for_each_desc(desc, dev, MSI_DESC_ALL) {
> +			if (desc == failed_desc)
> +				break;
> +			kfree(free_irq(desc->irq, epf));

All instances of interrupts get 'epf' as device id. So if the third
instance failed to be requested, you free 'epf' when freeing the first
interrupt and then again when freeing the second one.

Thanks,

        tglx