Hi,
On Wed, Mar 17, 2021 at 05:49:33PM -0700, Prashant Malani wrote:
> Hi Gil,
>
>
> On Tue, Mar 9, 2021 at 1:23 AM Gil Fine <gil.fine@xxxxxxxxx> wrote:
> >
> > Currently, when first failure occurs while reading of the block,
> > we stop reading the block and jump to the next capability.
> > This doesn't cover the case of block with "holes" of inaccessible
> > dwords, followed by accessible dwords.
> > This patch address this problem.
> > In case of failure while reading the complete block in one transaction,
> > (because of one or more dwords is inaccessible), we read the remaining
> > dwords of the block dword-by-dword, one dword per transaction,
> > till the end of the block.
> > By doing this, we handle the case of block with "holes" of inaccessible
> > dwords, followed by accessible dwords. The accessible dwords are shown
> > with the fields: <offset> <relative_offset> <cap_id> <vs_cap_id> <value>
> > E.g.:
> > 0x01eb 236 0x05 0x06 0x0000d166
> > While the inaccesible dwords are shown as: <offset> <not accessible>
> > E.g.:
> > 0x01ed <not accessible>
> >
> > Signed-off-by: Gil Fine <gil.fine@xxxxxxxxx>
> > ---
> > drivers/thunderbolt/debugfs.c | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/thunderbolt/debugfs.c b/drivers/thunderbolt/debugfs.c
> > index 201036507cb8..c850b0ac098c 100644
> > --- a/drivers/thunderbolt/debugfs.c
> > +++ b/drivers/thunderbolt/debugfs.c
> > @@ -265,10 +265,8 @@ static void cap_show_by_dw(struct seq_file *s, struct tb_switch *sw,
> > else
> > ret = tb_sw_read(sw, &data, TB_CFG_SWITCH, cap + offset + i, 1);
> > if (ret) {
> > - seq_printf(s, "0x%04x <not accessible>\n", cap + offset);
> > - if (dwords - i > 1)
> > - seq_printf(s, "0x%04x ...\n", cap + offset + 1);
> > - return;
> > + seq_printf(s, "0x%04x <not accessible>\n", cap + offset + i);
> > + continue;
> > }
> >
> > seq_printf(s, "0x%04x %4d 0x%02x 0x%02x 0x%08x\n", cap + offset + i,
> > @@ -292,7 +290,7 @@ static void cap_show(struct seq_file *s, struct tb_switch *sw,
> > else
> > ret = tb_sw_read(sw, data, TB_CFG_SWITCH, cap + offset, dwords);
> > if (ret) {
> > - cap_show_by_dw(s, sw, port, cap, offset, cap_id, vsec_id, dwords);
>
> Sorry for being late here:
> Can we call cap_show_by_dw(..., dwords) directly here, instead of
> having the logic of doing the block read and then resorting to this if
> the block read fails?
> Since it's debugfs, I doubt efficiency gains are paramount here, and
> we have a simpler invocation at the callsite.
I don't think that complicates too much, and in most cases we don't
expect the cap read to fail so doing block read saves us sending
multiple messages over the fabric (even if we don't care about the
performance here).
In any case the patch is already applied so if you want to change this
please send an incremental patch.
