linux-kernel - Re: [PATCH v10 1/2] crypto: hisilicon

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: <20251027222011.05bac6bd@shazbot.org>
Date: Mon, 27 Oct 2025 22:20:11 -0600
From: Alex Williamson <alex@...zbot.org>
To: Longfang Liu <liulongfang@...wei.com>
Cc: <alex.williamson@...hat.com>, <jgg@...dia.com>,
 <herbert@...dor.apana.org.au>, <shameerkolothum@...il.com>,
 <jonathan.cameron@...wei.com>, <linux-crypto@...r.kernel.org>,
 <kvm@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
 <linuxarm@...neuler.org>
Subject: Re: [PATCH v10 1/2] crypto: hisilicon - qm updates BAR
 configuration

On Fri, 17 Oct 2025 17:10:56 +0800
Longfang Liu <liulongfang@...wei.com> wrote:

> On new platforms greater than QM_HW_V3, the configuration region for the
> live migration function of the accelerator device is no longer
> placed in the VF, but is instead placed in the PF.
> 
> Therefore, the configuration region of the live migration function
> needs to be opened when the QM driver is loaded. When the QM driver
> is uninstalled, the driver needs to clear this configuration.
> 
> Signed-off-by: Longfang Liu <liulongfang@...wei.com>
> Reviewed-by: Shameer Kolothum <shameerkolothum@...il.com>
> Acked-by: Herbert Xu <herbert@...dor.apana.org.au>
> ---
>  drivers/crypto/hisilicon/qm.c | 27 +++++++++++++++++++++++++++
>  include/linux/hisi_acc_qm.h   |  3 +++
>  2 files changed, 30 insertions(+)
> 
> diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
> index a5b96adf2d1e..f0fd0c3698eb 100644
> --- a/drivers/crypto/hisilicon/qm.c
> +++ b/drivers/crypto/hisilicon/qm.c
> @@ -3019,11 +3019,36 @@ static void qm_put_pci_res(struct hisi_qm *qm)
>  	pci_release_mem_regions(pdev);
>  }
>  
> +static void hisi_mig_region_clear(struct hisi_qm *qm)
> +{
> +	u32 val;
> +
> +	/* Clear migration region set of PF */
> +	if (qm->fun_type == QM_HW_PF && qm->ver > QM_HW_V3) {
> +		val = readl(qm->io_base + QM_MIG_REGION_SEL);
> +		val &= ~BIT(0);
> +		writel(val, qm->io_base + QM_MIG_REGION_SEL);
> +	}
> +}
> +
> +static void hisi_mig_region_enable(struct hisi_qm *qm)
> +{
> +	u32 val;
> +
> +	/* Select migration region of PF */
> +	if (qm->fun_type == QM_HW_PF && qm->ver > QM_HW_V3) {
> +		val = readl(qm->io_base + QM_MIG_REGION_SEL);
> +		val |= QM_MIG_REGION_EN;
> +		writel(val, qm->io_base + QM_MIG_REGION_SEL);
> +	}
> +}

Why the inconsistency in using BIT(0) in one place and QM_MIG_REGION_EN
in another?  This could easily define QM_BIT_REGION_EN_BIT and use
BIT() in both places, or use the complement rather than BIT().  Thanks,

Alex

> +
>  static void hisi_qm_pci_uninit(struct hisi_qm *qm)
>  {
>  	struct pci_dev *pdev = qm->pdev;
>  
>  	pci_free_irq_vectors(pdev);
> +	hisi_mig_region_clear(qm);
>  	qm_put_pci_res(qm);
>  	pci_disable_device(pdev);
>  }
> @@ -5725,6 +5750,7 @@ int hisi_qm_init(struct hisi_qm *qm)
>  		goto err_free_qm_memory;
>  
>  	qm_cmd_init(qm);
> +	hisi_mig_region_enable(qm);
>  
>  	return 0;
>  
> @@ -5863,6 +5889,7 @@ static int qm_rebuild_for_resume(struct hisi_qm *qm)
>  	}
>  
>  	qm_cmd_init(qm);
> +	hisi_mig_region_enable(qm);
>  	hisi_qm_dev_err_init(qm);
>  	/* Set the doorbell timeout to QM_DB_TIMEOUT_CFG ns. */
>  	writel(QM_DB_TIMEOUT_SET, qm->io_base + QM_DB_TIMEOUT_CFG);
> diff --git a/include/linux/hisi_acc_qm.h b/include/linux/hisi_acc_qm.h
> index c4690e365ade..aa0129d20c51 100644
> --- a/include/linux/hisi_acc_qm.h
> +++ b/include/linux/hisi_acc_qm.h
> @@ -99,6 +99,9 @@
>  
>  #define QM_DEV_ALG_MAX_LEN		256
>  
> +#define QM_MIG_REGION_SEL		0x100198
> +#define QM_MIG_REGION_EN		0x1
> +
>  /* uacce mode of the driver */
>  #define UACCE_MODE_NOUACCE		0 /* don't use uacce */
>  #define UACCE_MODE_SVA			1 /* use uacce sva mode */