drm/xe/mmio: Avoid double-adjust in 64-bit reads

xe_mmio_read64_2x32() was adjusting register addresses and then
calling xe_mmio_read32(), which applies the adjustment again.
This may shift accesses twice if adj_offset < adj_limit. There is
no issue currently, as for media gt, adj_offset > adj_limit, so
the 2nd adjust will be a no-op. But it may not work in future.

To fix it, replace the adjusted-address comparison with a direct
sanity check that ensures the MMIO address adjustment cutoff never
falls within the 8-byte range of a 64-bit register. And let
xe_mmio_read32() handle address translation.

v2: rewrite the sanity check in a more natural way. (Matt)
v3: Add Fixes tag. (Jani)

Fixes: 07431945d8 ("drm/xe: Avoid 64-bit register reads")
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Shuicheng Lin <shuicheng.lin@intel.com>
Link: https://patch.msgid.link/20260130165621.471408-2-shuicheng.lin@intel.com
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
(cherry picked from commit a30f999681126b128a43137793ac84b6a5b7443f)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This commit is contained in:
Shuicheng Lin
2026-01-30 16:56:22 +00:00
committed by Rodrigo Vivi
parent fbbe32618e
commit 4a9b4e1fa5

View File

@@ -256,11 +256,11 @@ u64 xe_mmio_read64_2x32(struct xe_mmio *mmio, struct xe_reg reg)
struct xe_reg reg_udw = { .addr = reg.addr + 0x4 };
u32 ldw, udw, oldudw, retries;
reg.addr = xe_mmio_adjusted_addr(mmio, reg.addr);
reg_udw.addr = xe_mmio_adjusted_addr(mmio, reg_udw.addr);
/* we shouldn't adjust just one register address */
xe_tile_assert(mmio->tile, reg_udw.addr == reg.addr + 0x4);
/*
* The two dwords of a 64-bit register can never straddle the offset
* adjustment cutoff.
*/
xe_tile_assert(mmio->tile, !in_range(mmio->adj_limit, reg.addr + 1, 7));
oldudw = xe_mmio_read32(mmio, reg_udw);
for (retries = 5; retries; --retries) {