Thread (18 messages) flat view 18 messages, 8 authors, 2026-04-09

Re: [PATCH v2 4/4] drm/xe: switch xe_pagefault_queue_init() to using bitmap_weighted_or()

From: <anthony.l.nguyen@intel.com>
Date: 2026-03-30 20:48:13
Also in: dri-devel, intel-wired-lan, intel-xe, lkml


On 3/28/26 2:57 PM, Yury Norov [off-list ref] wrote:
On Thu, Mar 05, 2026 at 02:40:53PM -0800, Tony Nguyen wrote:
quoted

On 3/4/2026 3:43 AM, Simon Horman wrote:
quoted
On Sun, Mar 01, 2026 at 08:11:58PM -0500, Yury Norov wrote:
quoted
The function calls bitmap_or() immediately followed by bitmap_weight().
Switch to using the dedicated bitmap_weighted_or() and save one bitmap
traverse.

Signed-off-by: Yury Norov <redacted>
It's not entirely clear to me why this patch is included in a patchset
for the ice driver.

And it's also not clear to me why, but allmodconfigs - fo4 at least x86_32
and x86_64 - fail with this patch applied to net-next [1].

ERROR: modpost: "__bitmap_weighted_or" [drivers/gpu/drm/xe/xe.ko] undefined!

[1] 2b12ffb66955 ("net: mana: Trigger VF reset/recovery on health check failure due to HWC timeout")
I'm also seeing the same error and no apparent reason. Since this doesn't
seem dependent on the others, I'll take the other 3 and this can get
sent/resolved separately.
That's because the symbol is not exported, and the driver is build as
module.

It's already fixed in -next: 95d324fb1b484 ("bitmap: add test_zero_nbits()").
Let me know if you want me to send the fix as a separate patch in your
tree. Or I can take this patch in my branch, if you give me your tags.
IIRC there weren't any dependencies on this patch with the others, so I don't
see a need for this to go via the Intel tree. I believe this can submitted by itself?

Thanks,
Tony
Thanks,
Yury
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help