linux-kernel - Re: [PATCH] nvmem: core: Fix race in nvmem

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] [day] [month] [year] [list]

Message-ID: <6b7d1e08-e27e-3f8a-5b08-ba622f206295@linaro.org>
Date:   Tue, 3 Jan 2023 11:24:11 +0000
From:   Srinivas Kandagatla <srinivas.kandagatla@...aro.org>
To:     Hector Martin <marcan@...can.st>
Cc:     Sven Peter <sven@...npeter.dev>,
        Alyssa Rosenzweig <alyssa@...enzweig.io>,
        asahi@...ts.linux.dev, linux-kernel@...r.kernel.org,
        Russell King <rmk+kernel@...linux.org.uk>
Subject: Re: [PATCH] nvmem: core: Fix race in nvmem_register()

Thanks Hector for the patch,


On 14/12/2022 14:29, Hector Martin wrote:
> nvmem_register() currently registers the device before adding the nvmem
> cells, which creates a race window where consumers may find the nvmem
> device (and not get PROBE_DEFERred), but then fail to find the cells and
> error out.
> 
> Move device registration to the end of nvmem_register(), to close the
> race.
> 
> Observed when the stars line up on Apple Silicon machines with the (not
> yet upstream, but trivial) spmi nvmem driver and the macsmc-rtc client:
> 
> [    0.487375] macsmc-rtc macsmc-rtc: error -ENOENT: Failed to get rtc_offset NVMEM cell
> 
> Fixes: eace75cfdcf7 ("nvmem: Add a simple NVMEM framework for nvmem providers")
> Signed-off-by: Hector Martin <marcan@...can.st>

Missing Cc Stable

> ---
>   drivers/nvmem/core.c | 18 ++++++++----------
>   1 file changed, 8 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> index 321d7d63e068..d255feca5e17 100644
> --- a/drivers/nvmem/core.c
> +++ b/drivers/nvmem/core.c
> @@ -837,20 +837,16 @@ struct nvmem_device *nvmem_register(const struct nvmem_config *config)
> 
>   	dev_dbg(&nvmem->dev, "Registering nvmem device %s\n", config->name);
> 
> -	rval = device_register(&nvmem->dev);
> -	if (rval)
> -		goto err_put_device;
> -
>   	if (nvmem->nkeepout) {
>   		rval = nvmem_validate_keepouts(nvmem);
>   		if (rval)
> -			goto err_device_del;
> +			return ERR_PTR(rval);

you can not do this as this will leak ida and nvmem itself.


>   	}
> 
>   	if (config->compat) {
>   		rval = nvmem_sysfs_setup_compat(nvmem, config);
>   		if (rval)
> -			goto err_device_del;
> +			return ERR_PTR(rval);

same here.

>   	}
> 
>   	if (config->cells) {
> @@ -867,19 +863,21 @@ struct nvmem_device *nvmem_register(const struct nvmem_config *config)
>   	if (rval)
>   		goto err_remove_cells;
> 
> +	rval = device_register(&nvmem->dev);
> +	if (rval)
> +		goto err_put_device;
> +
>   	blocking_notifier_call_chain(&nvmem_notifier, NVMEM_ADD, nvmem);
> 
>   	return nvmem;
> 
> +err_put_device:
> +	put_device(&nvmem->dev);

this will free nvmem and further down you are going to de-reference the 
freed pointer.

--srini

>   err_remove_cells:
>   	nvmem_device_remove_all_cells(nvmem);
>   err_teardown_compat:
>   	if (config->compat)
>   		nvmem_sysfs_remove_compat(nvmem, config);
> -err_device_del:
> -	device_del(&nvmem->dev);
> -err_put_device:
> -	put_device(&nvmem->dev);
> 
>   	return ERR_PTR(rval);
>   }
> --
> 2.35.1
>