linux-kernel - Re: [PATCH] PCI: altera-msi: Remove irq handler and data in one go

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 for Android: free password hash cracker in your pocket

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

Date:   Tue, 10 Nov 2020 15:21:34 -0600
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Martin Kaiser <martin@...ser.cx>
Cc:     Ley Foon Tan <ley.foon.tan@...el.com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        rfi@...ts.rocketboards.org, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Nicolas Saenz Julienne <nsaenzjulienne@...e.de>,
        Jingoo Han <jingoohan1@...il.com>,
        Gustavo Pimentel <gustavo.pimentel@...opsys.com>,
        Toan Le <toan@...amperecomputing.com>
Subject: Re: [PATCH] PCI: altera-msi: Remove irq handler and data in one go

[+cc Nicolas, Jingoo, Gustavo, Toan]

On Sun, Nov 08, 2020 at 08:11:40PM +0100, Martin Kaiser wrote:
> Replace the two separate calls for removing the irq handler and data with a
> single irq_set_chained_handler_and_data() call.

This is similar to these:

  36f024ed8fc9 ("PCI/MSI: pci-xgene-msi: Consolidate chained IRQ handler install/remove")
  5168a73ce32d ("PCI/keystone: Consolidate chained IRQ handler install/remove")
  2cf5a03cb29d ("PCI/keystone: Fix race in installing chained IRQ handler")

and it seems potentially important that this removes the IRQ handler
and data *atomically*, i.e., both are done while holding
irq_get_desc_buslock().  

So I would use this:

  PCI: altera-msi: Fix race in installing chained IRQ handler

  Fix a race where a pending interrupt could be received and the handler
  called before the handler's data has been setup by converting to
  irq_set_chained_handler_and_data().

  See also 2cf5a03cb29d ("PCI/keystone: Fix race in installing chained
  IRQ handler").

to make it clear that this is actually a bug fix, not just a cleanup.

Looks like this should also be done in dw_pcie_free_msi() and
xgene_msi_hwirq_alloc() at the same time?

> Signed-off-by: Martin Kaiser <martin@...ser.cx>
> ---
>  drivers/pci/controller/pcie-altera-msi.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/controller/pcie-altera-msi.c b/drivers/pci/controller/pcie-altera-msi.c
> index e1636f7714ca..42691dd8ebef 100644
> --- a/drivers/pci/controller/pcie-altera-msi.c
> +++ b/drivers/pci/controller/pcie-altera-msi.c
> @@ -204,8 +204,7 @@ static int altera_msi_remove(struct platform_device *pdev)
>  	struct altera_msi *msi = platform_get_drvdata(pdev);
>  
>  	msi_writel(msi, 0, MSI_INTMASK);
> -	irq_set_chained_handler(msi->irq, NULL);
> -	irq_set_handler_data(msi->irq, NULL);
> +	irq_set_chained_handler_and_data(msi->irq, NULL, NULL);
>  
>  	altera_free_domains(msi);
>  
> -- 
> 2.20.1
>