On 29/10/26 07:45, Michael Tokarev wrote:
On 4/17/26 12:51, Christian Speich wrote:
This patch series improves the performance of read/write/erase operations
on sdcards.
This is done by increasing the maximum buffer size that is worked on.
From 1 byte (master) to 512 bytes (commit 1-3) to larger than 512
(adma commit).
...
Christian Speich (7):
hw/sd: Switch read/write primitive to buf+len
hw/sd/sd: Allow multi-byte read/write for generic paths
hw/sd/sd: Use multi-byte/block writes for block path
hw/sd/sdhci: Don't use bounce buffer for ADMA
hw/sd/sdcard: Add erase-blocks-as-zero option.
hw/sd/sdcard: Optimize erase blocks as zero.
hw/sd: Update trace events for buf+len data
3 changes from this series landed in the master branch for v11.1.0
release:
468fa450a7e0c hw/sd: Switch read/write primitive to buf+len
eff58d3d9681a hw/sd/sd: Allow multi-byte read/write for generic paths
d7f3ab9d5a186 hw/sd: Update trace events for buf+len data
I don't see discussion about the other 4 changes, why these weren't
applied at the time, but ok.
The rest failed harness test suite.
Currently I'm picking up another change:
b5424abaf8e9a hw/sd/sdcard: Keep high-capacity memory blocks at 512 bytes
to the current stable qemu series, and it touches the same area.
But the mentioned-above 3 changes are not included in 10.0 release,
while 10.0.x is an LTS stable series. And without the first change
(switch to buf+len), things in 10.0.x are quite different.
So I can either pick the above 3 changes for 10.0.x, or I can omit
b5424abaf8e9a for 10.0.x.
To me it feels like picking up the above 3 optimization changes is the
way to go for 10.0.x.
What do you think?
I'll defer that to Bin, no strong opinion.
BTW, doesn't b5424abaf8e9a actually fix 468fa450a7e0c, not 2d7adea4fe57?
Likely.
Thanks,
/mjt
