lkml.org 
[lkml]   [2021]   [Jul]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [Intel-wired-lan] [PATCH] driver core: fix e1000e ltr bug
From
Date
On 6/30/2021 09:26, Yee Li wrote:
> I agree with you.
> That's better code.
Thank Yee-Li for discovering this problem. I would suggest we (intel)
will process a patch to address this problem. Can I put you kindly as
"Suggested-by"?
>
> Neftin, Sasha <sasha.neftin@intel.com <mailto:sasha.neftin@intel.com>>
> 于2021年6月30日周三 下午2:13写道:
>
> On 6/29/2021 20:33, Yee Li wrote:
> >
> > Yes, 18874368ns > 3145728ns.
> > But, 0xe40 < 0x1003.
> I got you. I would agree, direct comparison is error-prone. (10M is
> impacted)
> I would suggest do not use convert function. lat_en should rather
> presented as lat_enc = scale x value:
> Introduce two u16 variables, u16 lat_enc_d and u16 max_ltr_enc_d;
>
> lat_enc_d = (lat_enc & 0x0x3ff) x (1U << 5*((max_ltr_enc & 0x1c00)
> >> 10))
>
> max_ltr_enc_d = (max_ltr_enc & 0x0x3ff) x (1U << 5*((max_ltr_enc &
> 0x1c00) >> 10))
>
> Then:
> if (lat_enc_d > max_ltr_enc_d)
>         lat_enc = max_ltr_enc;
> what do you think?
>
> >
> > So, the final lat_enc is 0xe40.
> > (Latency encoded is less than maximum LTR encoded by platform)
> >
> > Neftin, Sasha <sasha.neftin@intel.com
> <mailto:sasha.neftin@intel.com> <mailto:sasha.neftin@intel.com
> <mailto:sasha.neftin@intel.com>>>
> > 于 2021年6月29日周二 22:49写道:
> >
> >     On 6/29/2021 11:21, YeeLi wrote:
> >     Yeeli,
> >      > In e1000e driver, a PCIe-like device, the max
> snoop/no-snoop latency
> >      > is the upper limit.So, directly compare the size of
> lat_enc and
> >      > max_ltr_enc is incorrect.
> >      >
> >     why?
> >      >      In 1000Mbps, 0x8b9 < 0x1003, 189440 ns < 3145728 ns,
> correct.
> >      >
> >      >      In 100Mbps, 0xc3a < 0x1003, 1900544 ns < 3145728 ns,
> correct.
> >      >
> >      >      In 10Mbps, 0xe40 < 0x1003, 18874368 > 3145728, incorrect.
> >      >
> >     Platform LTR encoded is 0x1003 - right. It is meant 1048576ns
> x 3 =
> >     3145738ns.
> >     Now,
> >     for 1000M: 0x08b9 => 185ns x 1024 = 189440ns (you are correct)
> >     for 100M: 0x0c3a => 58ns x 32768 = 1900544ns (correct)
> >     for 10M: 0x0e41 => 577ns x 32768 = 18907136ns (ok?)
> >     18907136ns > 3145738ns, (latency encoded is great than
> maximum LTR
> >     encoded by platform) - so, there is no point to wait more
> than platform
> >     required, and lat_enc=max_ltr_enc. It is expected and we sent
> right
> >     value to the power management controller.
> >     What is the problem you try solve?
> >
> >      > Decoded the lat_enc and max_ltr_enc before compare them is
> necessary.
> >      >
> >      > Signed-off-by: YeeLi <seven.yi.lee@gmail.com
> <mailto:seven.yi.lee@gmail.com>
> >     <mailto:seven.yi.lee@gmail.com <mailto:seven.yi.lee@gmail.com>>>
> >      > ---
> >      >   drivers/net/ethernet/intel/e1000e/ich8lan.c | 23
> >     ++++++++++++++++++++-
> >      >   1 file changed, 22 insertions(+), 1 deletion(-)
> >      >
> >      > diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c
> >     b/drivers/net/ethernet/intel/e1000e/ich8lan.c
> >      > index 590ad110d383..3bff1b570b76 100644
> >      > --- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
> >      > +++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
> >      > @@ -986,6 +986,27 @@ static s32
> e1000_k1_workaround_lpt_lp(struct
> >     e1000_hw *hw, bool link)
> >      >       return ret_val;
> >      >   }
> >      >
> >      > +static u32 convert_e1000e_ltr_scale(u32 val)
> >      > +{
> >      > +     if (val > 5)
> >      > +             return 0;
> >      > +
> >      > +     return 1U << (5 * val);
> >      > +}
> >      > +
> >      > +static u64 decoded_ltr(u32 val)
> >      > +{
> >      > +     u64 decoded_latency;
> >      > +     u32 value;
> >      > +     u32 scale;
> >      > +
> >      > +     value = val & 0x03FF;
> >      > +     scale = (val & 0x1C00) >> 10;
> >      > +     decoded_latency = value *
> convert_e1000e_ltr_scale(scale);
> >      > +
> >      > +     return decoded_latency;
> >      > +}
> >      > +
> >      >   /**
> >      >    *  e1000_platform_pm_pch_lpt - Set platform power
> management
> >     values
> >      >    *  @hw: pointer to the HW structure
> >      > @@ -1059,7 +1080,7 @@ static s32
> e1000_platform_pm_pch_lpt(struct
> >     e1000_hw *hw, bool link)
> >      >                                    E1000_PCI_LTR_CAP_LPT + 2,
> >     &max_nosnoop);
> >      >               max_ltr_enc = max_t(u16, max_snoop,
> max_nosnoop);
> >      >
> >      > -             if (lat_enc > max_ltr_enc)
> >      > +             if (decoded_ltr(lat_enc) >
> decoded_ltr(max_ltr_enc))
> >      >                       lat_enc = max_ltr_enc;
> >      >       }
> >      >
> >      >
> >     sasha
> >
>
sasha

\
 
 \ /
  Last update: 2021-07-01 10:35    [W:0.042 / U:0.676 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site