> From: hch@xxxxxx <hch@xxxxxx>
> Sent: Tuesday, October 27, 2020 1:41 PM
>
> On Mon, Oct 26, 2020 at 05:23:48AM +0000, Parav Pandit wrote:
> > Hi Christoph,
> >
> > > From: Jakub Kicinski <kuba@xxxxxxxxxx>
> > > Sent: Saturday, October 24, 2020 11:45 PM
> > >
> > > CC: rdma, looks like rdma from the stack trace
> > >
> > > On Fri, 23 Oct 2020 20:07:17 -0700 syzbot wrote:
> > > > syzbot has found a reproducer for the following issue on:
> > > >
> > > > HEAD commit: 3cb12d27 Merge tag 'net-5.10-rc1' of
> git://git.kernel.org/..
> >
> > In [1] you mentioned that dma_mask should not be set for dma_virt_ops.
> > So patch [2] removed it.
> >
> > But check to validate the dma mask for all dma_ops was added in [3].
> >
> > What is the right way? Did I misunderstood your comment about
> dma_mask in [1]?
>
> No, I did not say we don't need the mask. I said copying over the various
> dma-related fields from the parent is bogus.
>
> I think rxe (and ther other drivers/infiniband/sw drivers) need a simple
> dma_coerce_mask_and_coherent and nothing else.
I see. Does below fix make sense?
Is DMA_MASK_NONE correct?
>From cfad78c35788b4ff604abedd96559500c5fd2a72 Mon Sep 17 00:00:00 2001
From: Parav Pandit <parav@xxxxxxxxxx>
Date: Tue, 27 Oct 2020 14:20:07 +0200
Subject: [PATCH] RDMA: Fix software RDMA drivers for dma mapping error
A cited commit in fixes tag avoided setting dma_mask of the ib_device.
Commit [1] made dma_mask as mandetory field to be setup even for
dma_virt_ops based dma devices.
Fix it by setting empty DMA MASK for software based RDMA devices.
[1] commit: f959dcd6ddfd2 ("dma-direct: Fix potential NULL pointer dereference")
Reported-by: syzbot+34dc2fea3478e659af01@xxxxxxxxxxxxxxxxxxxxxxxxx
Fixes: e0477b34d9d1 ("RDMA: Explicitly pass in the dma_device to ib_register_device")
Signed-off-by: Parav Pandit <parav@xxxxxxxxxx>
---
drivers/infiniband/sw/rdmavt/vt.c | 5 +++--
drivers/infiniband/sw/rxe/rxe_verbs.c | 4 +++-
drivers/infiniband/sw/siw/siw_main.c | 5 +++--
3 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/drivers/infiniband/sw/rdmavt/vt.c b/drivers/infiniband/sw/rdmavt/vt.c
index 52218684ad4a..1b456f4d4fcf 100644
--- a/drivers/infiniband/sw/rdmavt/vt.c
+++ b/drivers/infiniband/sw/rdmavt/vt.c
@@ -580,8 +580,9 @@ int rvt_register_device(struct rvt_dev_info *rdi)
/* DMA Operations */
rdi->ibdev.dev.dma_parms = rdi->ibdev.dev.parent->dma_parms;
- dma_set_coherent_mask(&rdi->ibdev.dev,
- rdi->ibdev.dev.parent->coherent_dma_mask);
+ ret = dma_coerce_mask_and_coherent(&rdi->ibdev.dev, DMA_MASK_NONE);
+ if (ret)
+ goto bail_wss;
/* Protection Domain */
spin_lock_init(&rdi->n_pds_lock);
diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
index 1fc022362fbe..357787688293 100644
--- a/drivers/infiniband/sw/rxe/rxe_verbs.c
+++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
@@ -1130,7 +1130,9 @@ int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name)
rxe->ndev->dev_addr);
dev->dev.dma_parms = &rxe->dma_parms;
dma_set_max_seg_size(&dev->dev, UINT_MAX);
- dma_set_coherent_mask(&dev->dev, dma_get_required_mask(&dev->dev));
+ err = dma_coerce_mask_and_coherent(&dev->dev, DMA_MASK_NONE);
+ if (err)
+ return err;
dev->uverbs_cmd_mask = BIT_ULL(IB_USER_VERBS_CMD_GET_CONTEXT)
| BIT_ULL(IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)
diff --git a/drivers/infiniband/sw/siw/siw_main.c b/drivers/infiniband/sw/siw/siw_main.c
index ca8bc7296867..d3dc50a42dab 100644
--- a/drivers/infiniband/sw/siw/siw_main.c
+++ b/drivers/infiniband/sw/siw/siw_main.c
@@ -384,8 +384,9 @@ static struct siw_device *siw_device_create(struct net_device *netdev)
base_dev->dev.parent = parent;
base_dev->dev.dma_parms = &sdev->dma_parms;
dma_set_max_seg_size(&base_dev->dev, UINT_MAX);
- dma_set_coherent_mask(&base_dev->dev,
- dma_get_required_mask(&base_dev->dev));
+ if (dma_coerce_mask_and_coherent(&base_dev->dev, DMA_MASK_NONE))
+ goto error;
+
base_dev->num_comp_vectors = num_possible_cpus();
xa_init_flags(&sdev->qp_xa, XA_FLAGS_ALLOC1);
--
2.26.2
