| 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
| ||
|
Message-ID: <YR0SEwGzgQaKAHYC@google.com>
Date: Wed, 18 Aug 2021 06:58:43 -0700
From: Matthias Kaehlcke <mka@...omium.org>
To: Zijun Hu <zijuhu@...eaurora.org>
Cc: marcel@...tmann.org, johan.hedberg@...il.com, luiz.dentz@...il.com,
linux-kernel@...r.kernel.org, linux-bluetooth@...r.kernel.org,
linux-arm-msm@...r.kernel.org, bgodavar@...eaurora.org,
c-hbandi@...eaurora.org, hemantg@...eaurora.org,
rjliao@...eaurora.org, tjiang@...eaurora.org
Subject: Re: [PATCH v3] Bluetooth: btusb: Add support different nvm to
distinguish different factory for WCN6855 controller
On Thu, Aug 12, 2021 at 04:50:16PM +0800, Zijun Hu wrote:
> From: Tim Jiang <tjiang@...eaurora.org>
>
> we have different factory to produce wcn6855 soc chip, so we should
> use different nvm file with suffix to distinguish them.
What exactly does factory mean in this context, different production
facilities? Could this be just treated as a variant of the chip instead
of introducing the concept of 'factory'?
>
> Signed-off-by: Tim Jiang <tjiang@...eaurora.org>
> ---
> drivers/bluetooth/btusb.c | 60 ++++++++++++++++++++++++++++++++++++++++-------
> 1 file changed, 51 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> index b1a05bb9f4bf..d7b4e0f1c3e3 100644
> --- a/drivers/bluetooth/btusb.c
> +++ b/drivers/bluetooth/btusb.c
> @@ -4013,6 +4013,9 @@ static int btusb_set_bdaddr_wcn6855(struct hci_dev *hdev,
> #define QCA_DFU_TIMEOUT 3000
> #define QCA_FLAG_MULTI_NVM 0x80
>
> +#define WCN6855_2_0_RAM_VERSION_GF 0x400c1200
> +#define WCN6855_2_1_RAM_VERSION_GF 0x400c1211
> +
> struct qca_version {
> __le32 rom_version;
> __le32 patch_version;
> @@ -4044,6 +4047,7 @@ static const struct qca_device_info qca_devices_table[] = {
> { 0x00000302, 28, 4, 16 }, /* Rome 3.2 */
> { 0x00130100, 40, 4, 16 }, /* WCN6855 1.0 */
> { 0x00130200, 40, 4, 16 }, /* WCN6855 2.0 */
> + { 0x00130201, 40, 4, 16 }, /* WCN6855 2.1 */
> };
>
> static int btusb_qca_send_vendor_req(struct usb_device *udev, u8 request,
> @@ -4198,6 +4202,39 @@ static int btusb_setup_qca_load_rampatch(struct hci_dev *hdev,
> return err;
> }
>
> +static int btusb_setup_qca_form_nvm_name(char **fwname,
How about btusb_generate/get_qca_fw_name()?
> + int max_size,
> + struct qca_version *ver,
> + char *factory)
> +{
> + /* if boardid equal 0, use default nvm without suffix */
> + switch (le16_to_cpu(ver->board_id)) {
> + case 0x0:
> + /* we add suffix factory to distinguish with different factory. */
> + if (factory != NULL) {
> + snprintf(*fwname, max_size, "qca/nvm_usb_%08x_%s.bin",
> + le32_to_cpu(ver->rom_version),
> + factory);
> + } else {
> + snprintf(*fwname, max_size, "qca/nvm_usb_%08x.bin",
> + le32_to_cpu(ver->rom_version));
> + }
how about:
snprintf(*fwname, max_size, "qca/nvm_usb_%08x%s.bin",
le32_to_cpu(ver->rom_version),
factory);
And you either pass the 'factory' including the underscore to the function, or
an empty string (potentially with a suitable define). That would eliminate the
need for the if/else construct here and below.
Or alternatively:
snprintf(*fwname, max_size, "qca/nvm_usb_%08x%s%s.bin",
le32_to_cpu(ver->rom_version),
separator, factory);
With separator defaulting to an empty string and being assigned to '_' when
'factory' is set.
> + break;
> + default:
> + if (factory != NULL) {
> + snprintf(*fwname, max_size, "qca/nvm_usb_%08x_%s_%04x.bin",
> + le32_to_cpu(ver->rom_version),
> + factory,
> + le16_to_cpu(ver->board_id));
> + } else {
> + snprintf(*fwname, max_size, "qca/nvm_usb_%08x_%04x.bin",
> + le32_to_cpu(ver->rom_version),
> + le16_to_cpu(ver->board_id));
> + }
> + break;
> + }
> +}
In case you keep the if/else constructs you should probably use local
variables to do the conversion of rom version and board id from LE to CPU
format once, instead of doing it 4 times (in terms of code, not execution).
> +
> static int btusb_setup_qca_load_nvm(struct hci_dev *hdev,
> struct qca_version *ver,
> const struct qca_device_info *info)
> @@ -4206,19 +4243,24 @@ static int btusb_setup_qca_load_nvm(struct hci_dev *hdev,
> char fwname[64];
> int err;
>
> - if (((ver->flag >> 8) & 0xff) == QCA_FLAG_MULTI_NVM) {
> - /* if boardid equal 0, use default nvm without surfix */
> - if (le16_to_cpu(ver->board_id) == 0x0) {
> + switch (ver->ram_version) {
> + case WCN6855_2_0_RAM_VERSION_GF:
> + case WCN6855_2_1_RAM_VERSION_GF:
> + if (((ver->flag >> 8) & 0xff) == QCA_FLAG_MULTI_NVM) {
> + btusb_setup_qca_form_nvm_name(&fwname, sizeof(fwname), ver, "gf");
> + } else {
> snprintf(fwname, sizeof(fwname), "qca/nvm_usb_%08x.bin",
> le32_to_cpu(ver->rom_version));
> + }
The addition of a function to generate the firmware names makes sense IMO if
that gets more complex, to separate it from the actual firmware loading. However
it seems odd to only outsource part of it. Wouldn't it make more sense to hide
the entire complexity in btusb_setup_qca_form_nvm_name()?
> + break;
> + default:
> + if (((ver->flag >> 8) & 0xff) == QCA_FLAG_MULTI_NVM) {
> + btusb_setup_qca_form_nvm_name(&fwname, sizeof(fwname), ver, NULL);
> } else {
> - snprintf(fwname, sizeof(fwname), "qca/nvm_usb_%08x_%04x.bin",
> - le32_to_cpu(ver->rom_version),
> - le16_to_cpu(ver->board_id));
> + snprintf(fwname, sizeof(fwname), "qca/nvm_usb_%08x.bin",
> + le32_to_cpu(ver->rom_version));
> }
> - } else {
> - snprintf(fwname, sizeof(fwname), "qca/nvm_usb_%08x.bin",
> - le32_to_cpu(ver->rom_version));
> + break;
> }
>
> err = request_firmware(&fw, fwname, &hdev->dev);
> --
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
>
Powered by blists - more mailing lists