From: "Christian König" <christian.koenig@amd.com>
To: "Chang, HaiJun" <HaiJun.Chang@amd.com>,
"Shi, Lianjie" <Lianjie.Shi@amd.com>,
"linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Bjorn Helgaas <bhelgaas@google.com>
Cc: "Jiang, Jerry (SW)" <Jerry.Jiang@amd.com>,
"Zhang, Andy" <Andy.Zhang@amd.com>
Subject: Re: [PATCH 1/1][v2] PCI: Support VF resizable BAR
Date: Fri, 7 Jun 2024 13:37:21 +0200 [thread overview]
Message-ID: <432d2df8-3686-4f87-af1a-a9a172d662b9@amd.com> (raw)
In-Reply-To: <DM6PR12MB49584B9866498A33D79EFEC881FB2@DM6PR12MB4958.namprd12.prod.outlook.com>
I need the original set of patches, e.g. directly use git send-email
with my mail address.
Forwarded and mangled multiple times the patches are not applicable to
any tree any more.
Regards,
Christian.
Am 07.06.24 um 10:42 schrieb Chang, HaiJun:
> [AMD Official Use Only - AMD Internal Distribution Only]
>
> + Christian to review the VF resizable BAR patch.
>
> -----Original Message-----
> From: Chang, HaiJun
> Sent: Friday, May 31, 2024 12:26 PM
> To: Lianjie Shi <Lianjie.Shi@amd.com>; linux-pci@vger.kernel.org; linux-kernel@vger.kernel.org; Bjorn Helgaas <bhelgaas@google.com>
> Cc: Jiang, Jerry (SW) <Jerry.Jiang@amd.com>; Zhang, Andy <Andy.Zhang@amd.com>; Shi, Lianjie <Lianjie.Shi@amd.com>
> Subject: RE: [PATCH 1/1][v2] PCI: Support VF resizable BAR
>
> Hi, Bjorn and pci driver maintainers
>
> Could you review the VF resizable BAR patch?
>
> Thanks,
> HaiJun
>
> -----Original Message-----
> From: Lianjie Shi <Lianjie.Shi@amd.com>
> Sent: Thursday, May 23, 2024 3:11 PM
> To: linux-pci@vger.kernel.org; linux-kernel@vger.kernel.org; Bjorn Helgaas <bhelgaas@google.com>
> Cc: Chang, HaiJun <HaiJun.Chang@amd.com>; Jiang, Jerry (SW) <Jerry.Jiang@amd.com>; Zhang, Andy <Andy.Zhang@amd.com>; Shi, Lianjie <Lianjie.Shi@amd.com>
> Subject: [PATCH 1/1][v2] PCI: Support VF resizable BAR
>
> Add support for VF resizable BAR PCI extended cap.
> Similar to regular BAR, drivers can use pci_resize_resource() to resize an IOV BAR. For each VF, dev->sriov->barsz of the IOV BAR is resized, but the total resource size of the IOV resource should not exceed its original size upon init.
>
> Based on following patch series:
> Link: https://lore.kernel.org/lkml/YbqGplTKl5i%2F1%2FkY@rocinante/T/
>
> Signed-off-by: Lianjie Shi <Lianjie.Shi@amd.com>
> ---
> drivers/pci/pci.c | 47 ++++++++++++++++++++++++++++++++++-
> drivers/pci/setup-res.c | 45 +++++++++++++++++++++++++++------
> include/uapi/linux/pci_regs.h | 1 +
> 3 files changed, 85 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index e5f243dd4..12f86e00a 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -1867,6 +1867,42 @@ static void pci_restore_rebar_state(struct pci_dev *pdev)
> }
> }
>
> +static void pci_restore_vf_rebar_state(struct pci_dev *pdev) { #ifdef
> +CONFIG_PCI_IOV
> + unsigned int pos, nbars, i;
> + u32 ctrl;
> + u16 total;
> +
> + pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_VF_REBAR);
> + if (!pos)
> + return;
> +
> + pci_read_config_dword(pdev, pos + PCI_REBAR_CTRL, &ctrl);
> + nbars = FIELD_GET(PCI_REBAR_CTRL_NBAR_MASK, ctrl);
> +
> + for (i = 0; i < nbars; i++, pos += 8) {
> + struct resource *res;
> + int bar_idx, size;
> + u64 tmp;
> +
> + pci_read_config_dword(pdev, pos + PCI_REBAR_CTRL, &ctrl);
> + bar_idx = ctrl & PCI_REBAR_CTRL_BAR_IDX;
> + total = pdev->sriov->total_VFs;
> + if (!total)
> + return;
> +
> + res = pdev->resource + bar_idx + PCI_IOV_RESOURCES;
> + tmp = resource_size(res);
> + do_div(tmp, total);
> + size = pci_rebar_bytes_to_size(tmp);
> + ctrl &= ~PCI_REBAR_CTRL_BAR_SIZE;
> + ctrl |= FIELD_PREP(PCI_REBAR_CTRL_BAR_SIZE, size);
> + pci_write_config_dword(pdev, pos + PCI_REBAR_CTRL, ctrl);
> + }
> +#endif
> +}
> +
> /**
> * pci_restore_state - Restore the saved state of a PCI device
> * @dev: PCI device that we're dealing with @@ -1882,6 +1918,7 @@ void pci_restore_state(struct pci_dev *dev)
> pci_restore_ats_state(dev);
> pci_restore_vc_state(dev);
> pci_restore_rebar_state(dev);
> + pci_restore_vf_rebar_state(dev);
> pci_restore_dpc_state(dev);
> pci_restore_ptm_state(dev);
>
> @@ -3677,10 +3714,18 @@ void pci_acs_init(struct pci_dev *dev)
> */
> static int pci_rebar_find_pos(struct pci_dev *pdev, int bar) {
> + int cap = PCI_EXT_CAP_ID_REBAR;
> unsigned int pos, nbars, i;
> u32 ctrl;
>
> - pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_REBAR);
> +#ifdef CONFIG_PCI_IOV
> + if (bar >= PCI_IOV_RESOURCES) {
> + cap = PCI_EXT_CAP_ID_VF_REBAR;
> + bar -= PCI_IOV_RESOURCES;
> + }
> +#endif
> +
> + pos = pci_find_ext_capability(pdev, cap);
> if (!pos)
> return -ENOTSUPP;
>
> diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c index c6d933ddf..d978a2ccf 100644
> --- a/drivers/pci/setup-res.c
> +++ b/drivers/pci/setup-res.c
> @@ -427,13 +427,32 @@ void pci_release_resource(struct pci_dev *dev, int resno) } EXPORT_SYMBOL(pci_release_resource);
>
> +static int pci_memory_decoding(struct pci_dev *dev, int resno) {
> + u16 cmd;
> +
> +#ifdef CONFIG_PCI_IOV
> + if (resno >= PCI_IOV_RESOURCES) {
> + pci_read_config_word(dev, dev->sriov->pos + PCI_SRIOV_CTRL, &cmd);
> + if (cmd & PCI_SRIOV_CTRL_MSE)
> + return -EBUSY;
> + else
> + return 0;
> + }
> +#endif
> + pci_read_config_word(dev, PCI_COMMAND, &cmd);
> + if (cmd & PCI_COMMAND_MEMORY)
> + return -EBUSY;
> +
> + return 0;
> +}
> +
> int pci_resize_resource(struct pci_dev *dev, int resno, int size) {
> struct resource *res = dev->resource + resno;
> struct pci_host_bridge *host;
> int old, ret;
> u32 sizes;
> - u16 cmd;
>
> /* Check if we must preserve the firmware's resource assignment */
> host = pci_find_host_bridge(dev->bus); @@ -444,9 +463,9 @@ int pci_resize_resource(struct pci_dev *dev, int resno, int size)
> if (!(res->flags & IORESOURCE_UNSET))
> return -EBUSY;
>
> - pci_read_config_word(dev, PCI_COMMAND, &cmd);
> - if (cmd & PCI_COMMAND_MEMORY)
> - return -EBUSY;
> + ret = pci_memory_decoding(dev, resno);
> + if (ret)
> + return ret;
>
> sizes = pci_rebar_get_possible_sizes(dev, resno);
> if (!sizes)
> @@ -463,19 +482,31 @@ int pci_resize_resource(struct pci_dev *dev, int resno, int size)
> if (ret)
> return ret;
>
> - res->end = res->start + pci_rebar_size_to_bytes(size) - 1;
> +#ifdef CONFIG_PCI_IOV
> + if (resno >= PCI_IOV_RESOURCES)
> + dev->sriov->barsz[resno - PCI_IOV_RESOURCES] = pci_rebar_size_to_bytes(size);
> + else
> +#endif
> + res->end = res->start + pci_rebar_size_to_bytes(size) - 1;
>
> /* Check if the new config works by trying to assign everything. */
> if (dev->bus->self) {
> ret = pci_reassign_bridge_resources(dev->bus->self, res->flags);
> - if (ret)
> + if (ret && ret != -ENOENT)
> goto error_resize;
> }
> return 0;
>
> error_resize:
> pci_rebar_set_size(dev, resno, old);
> - res->end = res->start + pci_rebar_size_to_bytes(old) - 1;
> +
> +#ifdef CONFIG_PCI_IOV
> + if (resno >= PCI_IOV_RESOURCES)
> + dev->sriov->barsz[resno - PCI_IOV_RESOURCES] = pci_rebar_size_to_bytes(old);
> + else
> +#endif
> + res->end = res->start + pci_rebar_size_to_bytes(old) - 1;
> +
> return ret;
> }
> EXPORT_SYMBOL(pci_resize_resource);
> diff --git a/include/uapi/linux/pci_regs.h b/include/uapi/linux/pci_regs.h index a39193213..a66b90982 100644
> --- a/include/uapi/linux/pci_regs.h
> +++ b/include/uapi/linux/pci_regs.h
> @@ -738,6 +738,7 @@
> #define PCI_EXT_CAP_ID_L1SS 0x1E /* L1 PM Substates */
> #define PCI_EXT_CAP_ID_PTM 0x1F /* Precision Time Measurement */
> #define PCI_EXT_CAP_ID_DVSEC 0x23 /* Designated Vendor-Specific */
> +#define PCI_EXT_CAP_ID_VF_REBAR 0x24 /* VF Resizable BAR */
> #define PCI_EXT_CAP_ID_DLF 0x25 /* Data Link Feature */
> #define PCI_EXT_CAP_ID_PL_16GT 0x26 /* Physical Layer 16.0 GT/s */
> #define PCI_EXT_CAP_ID_PL_32GT 0x2A /* Physical Layer 32.0 GT/s */
> --
> 2.34.1
>
next prev parent reply other threads:[~2024-06-07 11:37 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-23 7:11 [PATCH 0/1][v2] PCI: VF resizable BAR Lianjie Shi
2024-05-23 7:11 ` [PATCH 1/1][v2] PCI: Support " Lianjie Shi
2024-05-31 4:26 ` Chang, HaiJun
2024-06-07 8:42 ` Chang, HaiJun
2024-06-07 11:37 ` Christian König [this message]
2024-06-07 20:40 ` Bjorn Helgaas
-- strict thread matches above, loose matches on Subject: below --
2024-06-12 8:42 [PATCH 0/1][v2] PCI: " Lianjie Shi
2024-06-12 8:42 ` [PATCH 1/1][v2] PCI: Support " Lianjie Shi
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=432d2df8-3686-4f87-af1a-a9a172d662b9@amd.com \
--to=christian.koenig@amd.com \
--cc=Andy.Zhang@amd.com \
--cc=HaiJun.Chang@amd.com \
--cc=Jerry.Jiang@amd.com \
--cc=Lianjie.Shi@amd.com \
--cc=bhelgaas@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox