All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jan Beulich <jbeulich@suse.com>
To: "xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>
Cc: "Kevin Tian" <kevin.tian@intel.com>,
	"Roger Pau Monné" <roger.pau@citrix.com>,
	"Paul Durrant" <paul@xen.org>
Subject: [PATCH v2 05/12] IOMMU: rename and re-type ats_enabled
Date: Thu, 15 Feb 2024 11:15:39 +0100	[thread overview]
Message-ID: <7f11ca06-9bed-443b-9c79-0e62b71a1f96@suse.com> (raw)
In-Reply-To: <64b028be-2197-4951-ae5b-32f9eabfa84a@suse.com>

Make the variable a tristate, with (as done elsewhere) a negative value
meaning "default". Since all use sites need looking at, also rename it
to match our usual "opt_*" pattern. While touching it, also move it to
.data.ro_after_init.

The only place it retains boolean nature is pci_ats_device(), for now.

In AMD code re-order conditionals to have the config space accesses
after (cheaper) flag checks.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
In domain_context_mapping_one() I'm a little puzzled that translation
type is selected based on only IOMMU and global properties, i.e. not
taking the device itself into account.
---
v2: Re-base over new earlier patches.

--- a/xen/drivers/passthrough/amd/iommu_cmd.c
+++ b/xen/drivers/passthrough/amd/iommu_cmd.c
@@ -282,7 +282,7 @@ void amd_iommu_flush_iotlb(u8 devfn, con
     struct amd_iommu *iommu;
     unsigned int req_id, queueid, maxpend;
 
-    if ( !ats_enabled )
+    if ( opt_ats <= 0 )
         return;
 
     if ( !pci_ats_enabled(pdev->seg, pdev->bus, pdev->devfn) )
@@ -340,7 +340,7 @@ static void _amd_iommu_flush_pages(struc
         flush_command_buffer(iommu, 0);
     }
 
-    if ( ats_enabled )
+    if ( opt_ats > 0 )
     {
         amd_iommu_flush_all_iotlbs(d, daddr, order);
 
--- a/xen/drivers/passthrough/amd/pci_amd_iommu.c
+++ b/xen/drivers/passthrough/amd/pci_amd_iommu.c
@@ -119,7 +119,7 @@ static bool use_ats(
     const struct amd_iommu *iommu,
     const struct ivrs_mappings *ivrs_dev)
 {
-    return !ivrs_dev->block_ats &&
+    return opt_ats > 0 && !ivrs_dev->block_ats &&
            iommu_has_cap(iommu, PCI_CAP_IOTLB_SHIFT) &&
            pci_ats_device(iommu->seg, pdev->bus, pdev->devfn);
 }
@@ -196,7 +196,7 @@ static int __must_check amd_iommu_setup_
         dte->sys_mgt = MASK_EXTR(ivrs_dev->device_flags, ACPI_IVHD_SYSTEM_MGMT);
 
         if ( use_ats(pdev, iommu, ivrs_dev) )
-            dte->i = ats_enabled;
+            dte->i = true;
 
         spin_unlock_irqrestore(&iommu->lock, flags);
 
@@ -257,7 +257,7 @@ static int __must_check amd_iommu_setup_
                                          ACPI_IVHD_SYSTEM_MGMT));
 
         if ( use_ats(pdev, iommu, ivrs_dev) )
-            ASSERT(dte->i == ats_enabled);
+            ASSERT(dte->i);
 
         spin_unlock_irqrestore(&iommu->lock, flags);
 
--- a/xen/drivers/passthrough/ats.c
+++ b/xen/drivers/passthrough/ats.c
@@ -18,8 +18,8 @@
 #include <xen/pci_regs.h>
 #include "ats.h"
 
-bool __read_mostly ats_enabled;
-boolean_param("ats", ats_enabled);
+int8_t __ro_after_init opt_ats = -1;
+boolean_param("ats", opt_ats);
 
 int enable_ats_device(struct pci_dev *pdev, struct list_head *ats_list)
 {
--- a/xen/drivers/passthrough/ats.h
+++ b/xen/drivers/passthrough/ats.h
@@ -22,7 +22,7 @@
 #define ATS_QUEUE_DEPTH_MASK     0x1f
 #define ATS_ENABLE               (1<<15)
 
-extern bool ats_enabled;
+extern int8_t opt_ats;
 
 int enable_ats_device(struct pci_dev *pdev, struct list_head *ats_list);
 void disable_ats_device(struct pci_dev *pdev);
@@ -43,7 +43,7 @@ static inline int pci_ats_enabled(int se
 
 static inline int pci_ats_device(int seg, int bus, int devfn)
 {
-    if ( !ats_enabled )
+    if ( !opt_ats )
         return 0;
 
     return pci_find_ext_capability(PCI_SBDF(seg, bus, devfn),
--- a/xen/drivers/passthrough/vtd/iommu.c
+++ b/xen/drivers/passthrough/vtd/iommu.c
@@ -1543,7 +1543,7 @@ int domain_context_mapping_one(
         }
 
         context_set_address_root(lctxt, root);
-        if ( ats_enabled && ecap_dev_iotlb(iommu->ecap) )
+        if ( opt_ats > 0 && ecap_dev_iotlb(iommu->ecap) )
             context_set_translation_type(lctxt, CONTEXT_TT_DEV_IOTLB);
         else
             context_set_translation_type(lctxt, CONTEXT_TT_MULTI_LEVEL);
--- a/xen/drivers/passthrough/vtd/x86/ats.c
+++ b/xen/drivers/passthrough/vtd/x86/ats.c
@@ -46,7 +46,7 @@ int ats_device(const struct pci_dev *pde
     struct acpi_drhd_unit *ats_drhd;
     unsigned int pos, expfl = 0;
 
-    if ( !ats_enabled || !iommu_qinval )
+    if ( opt_ats <= 0 || !iommu_qinval )
         return 0;
 
     if ( !ecap_queued_inval(drhd->iommu->ecap) ||



  parent reply	other threads:[~2024-02-15 10:15 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-15 10:11 [PATCH v2 00/12] VT-d: SATC handling; ATS: tidying Jan Beulich
2024-02-15 10:13 ` [PATCH v2 01/12] VT-d: correct ATS checking for root complex integrated devices Jan Beulich
2024-05-03 14:01   ` Roger Pau Monné
2024-02-15 10:14 ` [PATCH v2 02/12] VT-d: tidy error handling of RMRR parsing Jan Beulich
2024-05-06  9:12   ` Roger Pau Monné
2024-05-06  9:21     ` Jan Beulich
2024-05-06  9:26       ` Roger Pau Monné
2024-02-15 10:14 ` [PATCH v2 03/12] VT-d: parse ACPI "SoC Integrated Address Translation Cache Reporting Structure"s Jan Beulich
2024-05-06 10:29   ` Roger Pau Monné
2024-05-06 11:01     ` Jan Beulich
2024-05-06 11:09       ` Roger Pau Monné
2024-02-15 10:15 ` [PATCH v2 04/12] AMD/IOMMU: add helper to check whether ATS is to be used for a device Jan Beulich
2024-05-06 11:27   ` Roger Pau Monné
2024-02-15 10:15 ` Jan Beulich [this message]
2024-02-15 10:21   ` [PATCH v2 05/12] IOMMU: rename and re-type ats_enabled Jan Beulich
2024-05-06 12:42   ` Roger Pau Monné
2024-05-06 13:20     ` Jan Beulich
2024-05-06 13:53       ` Roger Pau Monné
2024-02-15 10:16 ` [PATCH v2 06/12] VT-d: respect ACPI SATC's ATC_REQUIRED flag Jan Beulich
2024-05-06 13:38   ` Roger Pau Monné
2024-02-15 10:16 ` [PATCH v2 07/12] VT-d: replace find_ats_dev_drhd() Jan Beulich
2024-02-15 10:17 ` [PATCH v2 08/12] VT-d: move ats_device() to the sole file it's used from Jan Beulich
2024-02-15 10:18 ` [PATCH v2 09/12] VT-d: move dev_invalidate_iotlb() " Jan Beulich
2024-02-15 10:18 ` [PATCH v2 10/12] VT-d: move {,un}map_vtd_domain_page() Jan Beulich
2024-02-15 10:18 ` [PATCH v2 11/12] VT-d: drop flush_dev_iotlb parameter from IOTLB flush hook Jan Beulich
2024-05-06 14:06   ` Roger Pau Monné
2024-02-15 10:19 ` [PATCH v2 12/12] PCI/ATS: tidy {en,dis}able_ats_device() a little Jan Beulich
2024-05-06 14:10   ` Roger Pau Monné

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=7f11ca06-9bed-443b-9c79-0e62b71a1f96@suse.com \
    --to=jbeulich@suse.com \
    --cc=kevin.tian@intel.com \
    --cc=paul@xen.org \
    --cc=roger.pau@citrix.com \
    --cc=xen-devel@lists.xenproject.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 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.