diff mbox series

[net-next] bnxt_en: Allow to set switchdev mode without existing VFs

Message ID 20230406130455.1155362-1-ivecera@redhat.com (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series [net-next] bnxt_en: Allow to set switchdev mode without existing VFs | expand

Checks

Context Check Description
netdev/series_format success Single patches do not need cover letters
netdev/tree_selection success Clearly marked for net-next
netdev/fixes_present success Fixes tag not required for -next series
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 19 this patch: 19
netdev/cc_maintainers success CCed 6 of 6 maintainers
netdev/build_clang success Errors and warnings before: 18 this patch: 18
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/deprecated_api success None detected
netdev/check_selftest success No net selftest shell script
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 19 this patch: 19
netdev/checkpatch warning WARNING: 'an user' may be misspelled - perhaps 'a user'?
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0

Commit Message

Ivan Vecera April 6, 2023, 1:04 p.m. UTC
Remove an inability of bnxt_en driver to set eswitch to switchdev
mode without existing VFs by:

1. Allow to set switchdev mode in bnxt_dl_eswitch_mode_set() so
   representors are created only when num_vfs > 0 otherwise just
   set bp->eswitch_mode
2. Do not automatically change bp->eswitch_mode during
   bnxt_vf_reps_create() and bnxt_vf_reps_destroy() calls so
   the eswitch mode is managed only by an user by devlink.
   Just set temporarily bp->eswitch_mode to legacy to avoid
   re-opening of representors during destroy.
3. Create representors in bnxt_sriov_enable() if current eswitch
   mode is switchdev one

Tested by this sequence:
1. Set PF interface up
2. Set PF's eswitch mode to switchdev
3. Created N VFs
4. Checked that N representors were created
5. Set eswitch mode to legacy
6. Checked that representors were deleted
7. Set eswitch mode back to switchdev
8. Checked that representros were re-created
9. Deleted all VFs
10. Checked that all representors were deleted as well
11. Checked that current eswitch mode is still switchdev

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
---
 .../net/ethernet/broadcom/bnxt/bnxt_sriov.c   | 16 ++++++++++
 drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c | 29 ++++++++++++-------
 drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.h |  6 ++++
 3 files changed, 41 insertions(+), 10 deletions(-)

Comments

Leon Romanovsky April 9, 2023, 9:02 a.m. UTC | #1
On Thu, Apr 06, 2023 at 03:04:55PM +0200, Ivan Vecera wrote:
> Remove an inability of bnxt_en driver to set eswitch to switchdev
> mode without existing VFs by:
> 
> 1. Allow to set switchdev mode in bnxt_dl_eswitch_mode_set() so
>    representors are created only when num_vfs > 0 otherwise just
>    set bp->eswitch_mode
> 2. Do not automatically change bp->eswitch_mode during
>    bnxt_vf_reps_create() and bnxt_vf_reps_destroy() calls so
>    the eswitch mode is managed only by an user by devlink.
>    Just set temporarily bp->eswitch_mode to legacy to avoid
>    re-opening of representors during destroy.
> 3. Create representors in bnxt_sriov_enable() if current eswitch
>    mode is switchdev one
> 
> Tested by this sequence:
> 1. Set PF interface up
> 2. Set PF's eswitch mode to switchdev
> 3. Created N VFs
> 4. Checked that N representors were created
> 5. Set eswitch mode to legacy
> 6. Checked that representors were deleted
> 7. Set eswitch mode back to switchdev
> 8. Checked that representros were re-created

Why do you think that this last item is the right behavior?
IMHO all configurations which were done after you switched mode
should be cleared and not recreated while toggling.

Thanks

> 9. Deleted all VFs
> 10. Checked that all representors were deleted as well
> 11. Checked that current eswitch mode is still switchdev
> 
> Signed-off-by: Ivan Vecera <ivecera@redhat.com>
> ---
Ivan Vecera April 11, 2023, 5:40 a.m. UTC | #2
On 09. 04. 23 11:02, Leon Romanovsky wrote:
> On Thu, Apr 06, 2023 at 03:04:55PM +0200, Ivan Vecera wrote:
>> Remove an inability of bnxt_en driver to set eswitch to switchdev
>> mode without existing VFs by:
>>
>> 1. Allow to set switchdev mode in bnxt_dl_eswitch_mode_set() so
>>     representors are created only when num_vfs > 0 otherwise just
>>     set bp->eswitch_mode
>> 2. Do not automatically change bp->eswitch_mode during
>>     bnxt_vf_reps_create() and bnxt_vf_reps_destroy() calls so
>>     the eswitch mode is managed only by an user by devlink.
>>     Just set temporarily bp->eswitch_mode to legacy to avoid
>>     re-opening of representors during destroy.
>> 3. Create representors in bnxt_sriov_enable() if current eswitch
>>     mode is switchdev one
>>
>> Tested by this sequence:
>> 1. Set PF interface up
>> 2. Set PF's eswitch mode to switchdev
>> 3. Created N VFs
>> 4. Checked that N representors were created
>> 5. Set eswitch mode to legacy
>> 6. Checked that representors were deleted
>> 7. Set eswitch mode back to switchdev
>> 8. Checked that representros were re-created
> 
> Why do you think that this last item is the right behavior?
> IMHO all configurations which were done after you switched mode
> should be cleared and not recreated while toggling.

No, I mean that if I switch back to switchdev mode then representors
should be created again for existing virtual functions not that
representors should be restored with their existing state...
So the point 8 should say:
"8. Checked that representors exist again for VFs"

Ivan
Paolo Abeni April 11, 2023, 10:22 a.m. UTC | #3
On Thu, 2023-04-06 at 15:04 +0200, Ivan Vecera wrote:
> Remove an inability of bnxt_en driver to set eswitch to switchdev
> mode without existing VFs by:
> 
> 1. Allow to set switchdev mode in bnxt_dl_eswitch_mode_set() so
>    representors are created only when num_vfs > 0 otherwise just
>    set bp->eswitch_mode
> 2. Do not automatically change bp->eswitch_mode during
>    bnxt_vf_reps_create() and bnxt_vf_reps_destroy() calls so
>    the eswitch mode is managed only by an user by devlink.
>    Just set temporarily bp->eswitch_mode to legacy to avoid
>    re-opening of representors during destroy.
> 3. Create representors in bnxt_sriov_enable() if current eswitch
>    mode is switchdev one
> 
> Tested by this sequence:
> 1. Set PF interface up
> 2. Set PF's eswitch mode to switchdev
> 3. Created N VFs
> 4. Checked that N representors were created
> 5. Set eswitch mode to legacy
> 6. Checked that representors were deleted
> 7. Set eswitch mode back to switchdev
> 8. Checked that representros were re-created

Could you please update the commit message and re-post?

Thanks!

Paolo
Ivan Vecera April 11, 2023, 12:02 p.m. UTC | #4
On 11. 04. 23 12:22, Paolo Abeni wrote:
> On Thu, 2023-04-06 at 15:04 +0200, Ivan Vecera wrote:
>> Remove an inability of bnxt_en driver to set eswitch to switchdev
>> mode without existing VFs by:
>>
>> 1. Allow to set switchdev mode in bnxt_dl_eswitch_mode_set() so
>>     representors are created only when num_vfs > 0 otherwise just
>>     set bp->eswitch_mode
>> 2. Do not automatically change bp->eswitch_mode during
>>     bnxt_vf_reps_create() and bnxt_vf_reps_destroy() calls so
>>     the eswitch mode is managed only by an user by devlink.
>>     Just set temporarily bp->eswitch_mode to legacy to avoid
>>     re-opening of representors during destroy.
>> 3. Create representors in bnxt_sriov_enable() if current eswitch
>>     mode is switchdev one
>>
>> Tested by this sequence:
>> 1. Set PF interface up
>> 2. Set PF's eswitch mode to switchdev
>> 3. Created N VFs
>> 4. Checked that N representors were created
>> 5. Set eswitch mode to legacy
>> 6. Checked that representors were deleted
>> 7. Set eswitch mode back to switchdev
>> 8. Checked that representros were re-created
> 
> Could you please update the commit message and re-post?
> 
> Thanks!

of course..

I.
diff mbox series

Patch

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
index 3ed3a2b3b3a9..dde327f2c57e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
@@ -825,8 +825,24 @@  static int bnxt_sriov_enable(struct bnxt *bp, int *num_vfs)
 	if (rc)
 		goto err_out2;
 
+	if (bp->eswitch_mode != DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return 0;
+
+	/* Create representors for VFs in switchdev mode */
+	devl_lock(bp->dl);
+	rc = bnxt_vf_reps_create(bp);
+	devl_unlock(bp->dl);
+	if (rc) {
+		netdev_info(bp->dev, "Cannot enable VFS as representors cannot be created\n");
+		goto err_out3;
+	}
+
 	return 0;
 
+err_out3:
+	/* Disable SR-IOV */
+	pci_disable_sriov(bp->pdev);
+
 err_out2:
 	/* Free the resources reserved for various VF's */
 	bnxt_hwrm_func_vf_resource_free(bp, *num_vfs);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c
index fcc65890820a..2f1a1f2d2157 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c
@@ -356,10 +356,15 @@  void bnxt_vf_reps_destroy(struct bnxt *bp)
 	/* un-publish cfa_code_map so that RX path can't see it anymore */
 	kfree(bp->cfa_code_map);
 	bp->cfa_code_map = NULL;
-	bp->eswitch_mode = DEVLINK_ESWITCH_MODE_LEGACY;
 
-	if (closed)
+	if (closed) {
+		/* Temporarily set legacy mode to avoid re-opening
+		 * representors and restore switchdev mode after that.
+		 */
+		bp->eswitch_mode = DEVLINK_ESWITCH_MODE_LEGACY;
 		bnxt_open_nic(bp, false, false);
+		bp->eswitch_mode = DEVLINK_ESWITCH_MODE_SWITCHDEV;
+	}
 	rtnl_unlock();
 
 	/* Need to call vf_reps_destroy() outside of rntl_lock
@@ -482,7 +487,7 @@  static void bnxt_vf_rep_netdev_init(struct bnxt *bp, struct bnxt_vf_rep *vf_rep,
 	dev->min_mtu = ETH_ZLEN;
 }
 
-static int bnxt_vf_reps_create(struct bnxt *bp)
+int bnxt_vf_reps_create(struct bnxt *bp)
 {
 	u16 *cfa_code_map = NULL, num_vfs = pci_num_vf(bp->pdev);
 	struct bnxt_vf_rep *vf_rep;
@@ -535,7 +540,6 @@  static int bnxt_vf_reps_create(struct bnxt *bp)
 
 	/* publish cfa_code_map only after all VF-reps have been initialized */
 	bp->cfa_code_map = cfa_code_map;
-	bp->eswitch_mode = DEVLINK_ESWITCH_MODE_SWITCHDEV;
 	netif_keep_dst(bp->dev);
 	return 0;
 
@@ -559,6 +563,7 @@  int bnxt_dl_eswitch_mode_set(struct devlink *devlink, u16 mode,
 			     struct netlink_ext_ack *extack)
 {
 	struct bnxt *bp = bnxt_get_bp_from_dl(devlink);
+	int ret = 0;
 
 	if (bp->eswitch_mode == mode) {
 		netdev_info(bp->dev, "already in %s eswitch mode\n",
@@ -570,7 +575,7 @@  int bnxt_dl_eswitch_mode_set(struct devlink *devlink, u16 mode,
 	switch (mode) {
 	case DEVLINK_ESWITCH_MODE_LEGACY:
 		bnxt_vf_reps_destroy(bp);
-		return 0;
+		break;
 
 	case DEVLINK_ESWITCH_MODE_SWITCHDEV:
 		if (bp->hwrm_spec_code < 0x10803) {
@@ -578,15 +583,19 @@  int bnxt_dl_eswitch_mode_set(struct devlink *devlink, u16 mode,
 			return -ENOTSUPP;
 		}
 
-		if (pci_num_vf(bp->pdev) == 0) {
-			netdev_info(bp->dev, "Enable VFs before setting switchdev mode\n");
-			return -EPERM;
-		}
-		return bnxt_vf_reps_create(bp);
+		/* Create representors for existing VFs */
+		if (pci_num_vf(bp->pdev) > 0)
+			ret = bnxt_vf_reps_create(bp);
+		break;
 
 	default:
 		return -EINVAL;
 	}
+
+	if (!ret)
+		bp->eswitch_mode = mode;
+
+	return ret;
 }
 
 #endif
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.h
index 5637a84884d7..33a965631d0b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.h
@@ -14,6 +14,7 @@ 
 
 #define	MAX_CFA_CODE			65536
 
+int bnxt_vf_reps_create(struct bnxt *bp);
 void bnxt_vf_reps_destroy(struct bnxt *bp);
 void bnxt_vf_reps_close(struct bnxt *bp);
 void bnxt_vf_reps_open(struct bnxt *bp);
@@ -37,6 +38,11 @@  int bnxt_dl_eswitch_mode_set(struct devlink *devlink, u16 mode,
 
 #else
 
+static inline int bnxt_vf_reps_create(struct bnxt *bp)
+{
+	return 0;
+}
+
 static inline void bnxt_vf_reps_close(struct bnxt *bp)
 {
 }