linux-kernel - Re: [PATCH 0/2] mmc: sdhci: add quirk to disable bounce buffer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CACRpkdYuq-Vq5-=DNKgY7a753fP28LikTyNDshCTCv0XtEmKUw@mail.gmail.com>
Date: Fri, 24 Oct 2025 10:32:50 +0200
From: Linus Walleij <linus.walleij@...aro.org>
To: Adrian Hunter <adrian.hunter@...el.com>
Cc: Michael Garofalo <officialtechflashyt@...il.com>, linux-kernel@...r.kernel.org, 
	linux-mmc@...r.kernel.org, ulf.hansson@...aro.org
Subject: Re: [PATCH 0/2] mmc: sdhci: add quirk to disable bounce buffer

On Fri, Oct 24, 2025 at 7:37 AM Adrian Hunter <adrian.hunter@...el.com> wrote:

> > These patches were mildly revealing, I see this near the start of the logs:
> > [   32.222988] b43-sdio mmc1:0001:1: Chip ID 14e4:4318

B43 WLAN eh? That driver is heavily bitrotted.

Anyways b43 does:

#define B43_SDIO_BLOCK_SIZE    64    /* rx fifo max size in bytes */

sdio_set_block_size(func, B43_SDIO_BLOCK_SIZE);

So that is an indication it does not want anything bigger than
that in it's FIFO.

It does this by setting blksize down to 64 and then sending
(probably) a train of sg_list:ed requests that the bound buffer
will merge into a contiguous request much bigger than 64
bytes to speed things up.

But I don't think the bounce buffer has proper SDIO minimum
block size awareness. It should simply step out of the way when
we do SDIO.

I will send a patch you can test.

Yours,
Linus Walleij