Re: [RESEND PATCH net-next] r8169: add module param for control of ASPM disable.

[David Miller]

From: Andrew Lunn <andrew@xxxxxxx>
Date: Wed, 14 Feb 2018 14:40:40 +0100

> linux/drivers$ grep -ir aspm * | grep MODULE_
> gpu/drm/amd/amdgpu/amdgpu_drv.c:MODULE_PARM_DESC(aspm, "ASPM support (1 = enable, 0 = disable, -1 = auto)");
> gpu/drm/radeon/radeon_drv.c:MODULE_PARM_DESC(aspm, "ASPM support (1 = enable, 0 = disable, -1 = auto)");
> infiniband/hw/hfi1/pcie.c:MODULE_PARM_DESC(aspm, "PCIe ASPM: 0: disable, 1: enable, 2: dynamic");
> net/wireless/realtek/rtlwifi/rtl8192ee/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8188ee/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8821ae/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8192se/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8192ce/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8723be/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8192de/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> net/wireless/realtek/rtlwifi/rtl8723ae/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> staging/rts5208/rtsx.c:MODULE_PARM_DESC(aspm_l0s_l1_en, "enable device aspm");
> staging/rtlwifi/rtl8822be/sw.c:MODULE_PARM_DESC(aspm, "Set to 1 to enable ASPM (default 1)\n");
> 
> This patch seems to have the exact opposite of everybody else already
> does.
> 
> Maybe you can follow the AMD example, and default to -1, since you are
> proposing to mostly have it enabled, but disabled in one case?

This is just another good reminder of why module parameters are a
terrible user experience for just about anything device configuration
related.

"Let's add knob X using a module parameter!"

Then a dozen or so other drivers copy the same thing with subtly, or
not so subtly, different behaviors, defaults, and semantics.

For users, this leads to misery.