From: Matthew Auld <matthew.auld@intel.com>
To: Michal Wajdeczko <michal.wajdeczko@intel.com>,
intel-xe@lists.freedesktop.org
Subject: Re: [PATCH v3 1/6] drm/xe/mmio: Verify MMIO is available
Date: Mon, 22 Jun 2026 16:39:32 +0100 [thread overview]
Message-ID: <c8d29b22-9d58-4630-add2-28951c681ced@intel.com> (raw)
In-Reply-To: <20260622132342.19600-2-michal.wajdeczko@intel.com>
On 22/06/2026 14:23, Michal Wajdeczko wrote:
> We shouldn't access device registers after the device was unplugged
> or the MMIO bar (GTTMMADR) was unmapped. Instead of relying on the
> NPD splat due to zeroed tile->mmio.regs, which might be unreliable
> anyway as not all xe_mmio structs are using that directly, add an
> explicit check during all xe_mmio read/write operations to test if
> xe->mmio.regs are still mapped and safely abort with WARN if not.
>
> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
> ---
> v1: https://patchwork.freedesktop.org/patch/728625/?series=167403&rev=1
> v2: https://patchwork.freedesktop.org/patch/731360/?series=168108&rev=1
> v3: use WARN and early abort to avoid Oops (sashiko)
> ---
> drivers/gpu/drm/xe/xe_mmio.c | 21 +++++++++++++++++++++
> 1 file changed, 21 insertions(+)
>
> diff --git a/drivers/gpu/drm/xe/xe_mmio.c b/drivers/gpu/drm/xe/xe_mmio.c
> index 78adb303b663..7e0cefcd16bd 100644
> --- a/drivers/gpu/drm/xe/xe_mmio.c
> +++ b/drivers/gpu/drm/xe/xe_mmio.c
> @@ -17,6 +17,7 @@
> #include "xe_device.h"
> #include "xe_gt_sriov_vf.h"
> #include "xe_sriov.h"
> +#include "xe_tile_printk.h"
> #include "xe_trace.h"
> #include "xe_wa.h"
>
> @@ -128,6 +129,11 @@ void xe_mmio_init(struct xe_mmio *mmio, struct xe_tile *tile, void __iomem *ptr,
> mmio->tile = tile;
> }
>
> +static bool mmio_available(struct xe_mmio *mmio)
> +{
> + return !xe_tile_WARN_ON_ONCE(mmio->tile, !mmio->tile->xe->mmio.regs);
> +}
> +
> static void mmio_flush_pending_writes(struct xe_mmio *mmio)
> {
> #define DUMMY_REG_OFFSET 0x130030
> @@ -146,6 +152,9 @@ u8 xe_mmio_read8(struct xe_mmio *mmio, struct xe_reg reg)
> u32 addr = xe_mmio_adjusted_addr(mmio, reg.addr);
> u8 val;
>
> + if (!mmio_available(mmio))
> + return 0;
> +
> mmio_flush_pending_writes(mmio);
>
> val = readb(mmio->regs + addr);
> @@ -158,6 +167,9 @@ void xe_mmio_write8(struct xe_mmio *mmio, struct xe_reg reg, u8 val)
> {
> u32 addr = xe_mmio_adjusted_addr(mmio, reg.addr);
>
> + if (!mmio_available(mmio))
> + return;
> +
> trace_xe_reg_rw(mmio, true, addr, val, sizeof(val));
>
> writeb(val, mmio->regs + addr);
> @@ -168,6 +180,9 @@ u16 xe_mmio_read16(struct xe_mmio *mmio, struct xe_reg reg)
> u32 addr = xe_mmio_adjusted_addr(mmio, reg.addr);
> u16 val;
>
> + if (!mmio_available(mmio))
> + return 0;
> +
> mmio_flush_pending_writes(mmio);
>
> val = readw(mmio->regs + addr);
> @@ -180,6 +195,9 @@ void xe_mmio_write32(struct xe_mmio *mmio, struct xe_reg reg, u32 val)
> {
> u32 addr = xe_mmio_adjusted_addr(mmio, reg.addr);
>
> + if (!mmio_available(mmio))
> + return;
> +
> trace_xe_reg_rw(mmio, true, addr, val, sizeof(val));
>
> if (!reg.vf && IS_SRIOV_VF(mmio->tile->xe))
> @@ -194,6 +212,9 @@ u32 xe_mmio_read32(struct xe_mmio *mmio, struct xe_reg reg)
> u32 addr = xe_mmio_adjusted_addr(mmio, reg.addr);
> u32 val;
>
> + if (!mmio_available(mmio))
> + return 0;
> +
> mmio_flush_pending_writes(mmio);
>
> if (!reg.vf && IS_SRIOV_VF(mmio->tile->xe))
next prev parent reply other threads:[~2026-06-22 15:39 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-22 13:23 [PATCH v3 0/6] drm/xe/mmio: Small improvements Michal Wajdeczko
2026-06-22 13:23 ` [PATCH v3 1/6] drm/xe/mmio: Verify MMIO is available Michal Wajdeczko
2026-06-22 15:39 ` Matthew Auld [this message]
2026-06-22 13:23 ` [PATCH v3 2/6] drm/xe/mmio: Map MMIO BAR using managed version of pci_iomap Michal Wajdeczko
2026-06-22 15:43 ` Matthew Auld
2026-06-22 13:23 ` [PATCH v3 3/6] drm/xe/mmio: Add check for minimal BAR size Michal Wajdeczko
2026-06-22 15:43 ` Matthew Auld
2026-06-22 13:23 ` [PATCH v3 4/6] drm/xe/mmio: Drop tiles_fini action Michal Wajdeczko
2026-06-22 15:44 ` Matthew Auld
2026-06-22 13:23 ` [PATCH v3 5/6] drm/xe/mmio: Check MMIO BAR size when initializing tiles Michal Wajdeczko
2026-06-22 13:23 ` [PATCH v3 6/6] drm/xe/mmio: Prefer tile-based WARN message Michal Wajdeczko
2026-06-23 9:13 ` Matthew Auld
2026-06-22 17:56 ` ✓ CI.KUnit: success for drm/xe/mmio: Small improvements (rev3) Patchwork
2026-06-22 18:52 ` ✓ Xe.CI.BAT: " Patchwork
2026-06-22 23:13 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-06-23 16:47 ` Michal Wajdeczko
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=c8d29b22-9d58-4630-add2-28951c681ced@intel.com \
--to=matthew.auld@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=michal.wajdeczko@intel.com \
/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