gateware.iostream.IOStreamer: fix bug for incorrect sampling DDR inputs This also adds a testcase to check the correct sampling time.

gateware.iostream.IOStreamer: fix bug for incorrect sampling of input Before this commit IOStreamer would sample its input 1 cycle late. This lowers latency by 1 clock cycle. This commit also adds a latency-agnostic testcase the verifies that SDR IOStreamer samples its inputs at the right point in time. This commit also fixes test_basic. This commit also fixes and simplifies the skid buffer test. Some behaviors of the skid buffer are not tested anymore because we're about to change those behaviors.

manual: revisions/revC3: U13 function description updated. U13 has same function description as U20 (was same as U21). manual: revisions/revC3: U13 function description was the same as description for U12 and U21. Should be the same description as for U20. Probably copy/paste error.

software: update `pdm.min.lock`.

manual: make the sidebar slightly wider. From Furo's `_layout.scss`: // Because CSS variables can't be used in media queries. The fact that this // makes the layout non-user-configurable is a good thing. This is of course bullshit. All it means is that whenever anyone upgrades Furo they have to manually build it with the following patch: $sidebar-width: 16em; and then manually put `furo.css` into `_static/styles/`. I don't like being told what (not) to do.

manual: revC3: Added hardware description documentation for Glasgow revC3 board.

manual: exempt GitHub from anchor checking as a whole.

applet.memory.25x: use QSPI controller instead of SPI. This makes it possible to use higher transfer rates, at the cost of WP# and HOLD# pins becoming mandatory.

applet.interface.qspi_controller: new applet.

gateware.qspi.QSPIController: new component. This component is a fully pipelined, zero wait states SPI, dual-SPI, and quad-SPI controller designed for high performance memory device access.