All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrew Jones <andrew.jones@linux.dev>
To: Gong Shuai <gong.shuai@sanechips.com.cn>
Cc: kvm@vger.kernel.org, kvm-riscv@lists.infradead.org,
	 linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	iommu@lists.linux.dev,  Anup Patel <anup@brainfault.org>,
	Atish Patra <atish.patra@linux.dev>,
	 Paul Walmsley <pjw@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	 Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	 Tomasz Jeznach <tomasz.jeznach@linux.dev>,
	"Joerg Roedel (AMD)" <joro@8bytes.org>,
	 Will Deacon <will@kernel.org>,
	Robin Murphy <robin.murphy@arm.com>,
	 Paolo Bonzini <pbonzini@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	 Jason Gunthorpe <jgg@ziepe.ca>,
	Kevin Tian <kevin.tian@intel.com>,
	 Andrew Jones <andrew.jones@oss.qualcomm.com>,
	Fangyu Yu <fangyu.yu@linux.alibaba.com>,
	 Tan Chi <tanchi25@mails.ucas.ac.cn>,
	Wu Fei <wu.fei9@sanechips.com.cn>,
	 Zhao Qingye <zhao.qingye@sanechips.com.cn>
Subject: Re: [RFC PATCH v3 09/10] iommu/riscv: Add riscv_iommu_attach_dev_nested
Date: Sat, 21 Nov 2026 16:48:51 +0100	[thread overview]
Message-ID: <awG5Q7pmntSme1A2@orel> (raw)
In-Reply-To: <20261119132744.396426-10-gong.shuai@sanechips.com.cn>

On Thu, Nov 19, 2026 at 09:27:43PM +0800, Gong Shuai wrote:
...
> +	if (riscv_iommu_bond_link(s2_parent, dev, nested_domain->viommu->gscid))
> +		return -ENOMEM;
> +
> +	riscv_iommu_iodir_update(iommu, dev, &dc);
> +	riscv_iommu_bond_unlink(info->domain, dev);
> +	info->domain = s2_parent;
> +

When moving between the S2 parent and a nested domain on it, the parent
briefly has two bonds for dev. bond_link() puts the new bond after the
first bond with the same GSCID, which may come before dev's old bond,
and since it's then dev's first bond, bond_unlink() removes it. dev is
left bonded with its old GSCID, so its new GSCID misses invalidations
once no other device's bond has it.

Passing the old domain to bond_unlink(), like I do in [1] for a
different reason, would let it remove the bond matching both dev and
old's GSCID.

[1] https://lore.kernel.org/all/20260928143113.49838-12-andrew.jones@oss.qualcomm.com/

Thanks,
drew

WARNING: multiple messages have this Message-ID (diff)
From: Andrew Jones <andrew.jones@linux.dev>
To: Gong Shuai <gong.shuai@sanechips.com.cn>
Cc: kvm@vger.kernel.org, kvm-riscv@lists.infradead.org,
	 linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	iommu@lists.linux.dev,  Anup Patel <anup@brainfault.org>,
	Atish Patra <atish.patra@linux.dev>,
	 Paul Walmsley <pjw@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	 Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	 Tomasz Jeznach <tomasz.jeznach@linux.dev>,
	"Joerg Roedel (AMD)" <joro@8bytes.org>,
	 Will Deacon <will@kernel.org>,
	Robin Murphy <robin.murphy@arm.com>,
	 Paolo Bonzini <pbonzini@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	 Jason Gunthorpe <jgg@ziepe.ca>,
	Kevin Tian <kevin.tian@intel.com>,
	 Andrew Jones <andrew.jones@oss.qualcomm.com>,
	Fangyu Yu <fangyu.yu@linux.alibaba.com>,
	 Tan Chi <tanchi25@mails.ucas.ac.cn>,
	Wu Fei <wu.fei9@sanechips.com.cn>,
	 Zhao Qingye <zhao.qingye@sanechips.com.cn>
Subject: Re: [RFC PATCH v3 09/10] iommu/riscv: Add riscv_iommu_attach_dev_nested
Date: Sat, 21 Nov 2026 16:48:51 +0100	[thread overview]
Message-ID: <awG5Q7pmntSme1A2@orel> (raw)
In-Reply-To: <20261119132744.396426-10-gong.shuai@sanechips.com.cn>

On Thu, Nov 19, 2026 at 09:27:43PM +0800, Gong Shuai wrote:
...
> +	if (riscv_iommu_bond_link(s2_parent, dev, nested_domain->viommu->gscid))
> +		return -ENOMEM;
> +
> +	riscv_iommu_iodir_update(iommu, dev, &dc);
> +	riscv_iommu_bond_unlink(info->domain, dev);
> +	info->domain = s2_parent;
> +

When moving between the S2 parent and a nested domain on it, the parent
briefly has two bonds for dev. bond_link() puts the new bond after the
first bond with the same GSCID, which may come before dev's old bond,
and since it's then dev's first bond, bond_unlink() removes it. dev is
left bonded with its old GSCID, so its new GSCID misses invalidations
once no other device's bond has it.

Passing the old domain to bond_unlink(), like I do in [1] for a
different reason, would let it remove the bond matching both dev and
old's GSCID.

[1] https://lore.kernel.org/all/20260928143113.49838-12-andrew.jones@oss.qualcomm.com/

Thanks,
drew

-- 
kvm-riscv mailing list
kvm-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kvm-riscv

WARNING: multiple messages have this Message-ID (diff)
From: Andrew Jones <andrew.jones@linux.dev>
To: Gong Shuai <gong.shuai@sanechips.com.cn>
Cc: kvm@vger.kernel.org, kvm-riscv@lists.infradead.org,
	 linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	iommu@lists.linux.dev,  Anup Patel <anup@brainfault.org>,
	Atish Patra <atish.patra@linux.dev>,
	 Paul Walmsley <pjw@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	 Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	 Tomasz Jeznach <tomasz.jeznach@linux.dev>,
	"Joerg Roedel (AMD)" <joro@8bytes.org>,
	 Will Deacon <will@kernel.org>,
	Robin Murphy <robin.murphy@arm.com>,
	 Paolo Bonzini <pbonzini@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	 Jason Gunthorpe <jgg@ziepe.ca>,
	Kevin Tian <kevin.tian@intel.com>,
	 Andrew Jones <andrew.jones@oss.qualcomm.com>,
	Fangyu Yu <fangyu.yu@linux.alibaba.com>,
	 Tan Chi <tanchi25@mails.ucas.ac.cn>,
	Wu Fei <wu.fei9@sanechips.com.cn>,
	 Zhao Qingye <zhao.qingye@sanechips.com.cn>
Subject: Re: [RFC PATCH v3 09/10] iommu/riscv: Add riscv_iommu_attach_dev_nested
Date: Sat, 21 Nov 2026 16:48:51 +0100	[thread overview]
Message-ID: <awG5Q7pmntSme1A2@orel> (raw)
In-Reply-To: <20261119132744.396426-10-gong.shuai@sanechips.com.cn>

On Thu, Nov 19, 2026 at 09:27:43PM +0800, Gong Shuai wrote:
...
> +	if (riscv_iommu_bond_link(s2_parent, dev, nested_domain->viommu->gscid))
> +		return -ENOMEM;
> +
> +	riscv_iommu_iodir_update(iommu, dev, &dc);
> +	riscv_iommu_bond_unlink(info->domain, dev);
> +	info->domain = s2_parent;
> +

When moving between the S2 parent and a nested domain on it, the parent
briefly has two bonds for dev. bond_link() puts the new bond after the
first bond with the same GSCID, which may come before dev's old bond,
and since it's then dev's first bond, bond_unlink() removes it. dev is
left bonded with its old GSCID, so its new GSCID misses invalidations
once no other device's bond has it.

Passing the old domain to bond_unlink(), like I do in [1] for a
different reason, would let it remove the bond matching both dev and
old's GSCID.

[1] https://lore.kernel.org/all/20260928143113.49838-12-andrew.jones@oss.qualcomm.com/

Thanks,
drew

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

  reply	other threads:[~2026-11-21 15:49 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19 13:27 [RFC PATCH v3 00/10] iommu/riscv: Add nested translation (vIOMMU) support Gong Shuai
2026-11-19 13:27 ` Gong Shuai
2026-11-19 13:27 ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 01/10] iommu/riscv: Use DC.tc.V as the device context valid bit Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 02/10] iommufd: Add RISC-V IOMMU hwpt and vIOMMU types Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 03/10] iommufd: Add RISC-V IOMMU cache invalidation data type Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 04/10] iommu/riscv: Add vIOMMU object support Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 05/10] iommu/riscv: Add nested domain allocation Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-21 12:37   ` Andrew Jones
2026-11-21 12:37     ` Andrew Jones
2026-11-21 12:37     ` Andrew Jones
2026-11-19 13:27 ` [RFC PATCH v3 06/10] iommu/riscv: Add vIOMMU cache invalidation Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 07/10] iommufd: Add RISC-V IOMMU hardware info type Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 08/10] iommu/riscv: Add riscv_iommu_hw_info Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27 ` [RFC PATCH v3 09/10] iommu/riscv: Add riscv_iommu_attach_dev_nested Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-21 15:48   ` Andrew Jones [this message]
2026-11-21 15:48     ` Andrew Jones
2026-11-21 15:48     ` Andrew Jones
2026-11-19 13:27 ` [RFC PATCH v3 10/10] iommu/riscv: Flush the right address space on bond unlink Gong Shuai
2026-11-19 13:27   ` Gong Shuai
2026-11-19 13:27   ` Gong Shuai

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=awG5Q7pmntSme1A2@orel \
    --to=andrew.jones@linux.dev \
    --cc=alex@ghiti.fr \
    --cc=andrew.jones@oss.qualcomm.com \
    --cc=anup@brainfault.org \
    --cc=aou@eecs.berkeley.edu \
    --cc=atish.patra@linux.dev \
    --cc=fangyu.yu@linux.alibaba.com \
    --cc=gong.shuai@sanechips.com.cn \
    --cc=iommu@lists.linux.dev \
    --cc=jgg@ziepe.ca \
    --cc=joro@8bytes.org \
    --cc=kevin.tian@intel.com \
    --cc=kvm-riscv@lists.infradead.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=palmer@dabbelt.com \
    --cc=pbonzini@redhat.com \
    --cc=pjw@kernel.org \
    --cc=robin.murphy@arm.com \
    --cc=seanjc@google.com \
    --cc=tanchi25@mails.ucas.ac.cn \
    --cc=tomasz.jeznach@linux.dev \
    --cc=will@kernel.org \
    --cc=wu.fei9@sanechips.com.cn \
    --cc=zhao.qingye@sanechips.com.cn \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.