v3.28.0: the batch CAP install

The bulk workflow: one .cap and one set of install parameters applied to
card after card - swap cards and each one installs as soon as it is
equipped.  The server owns the loop, so closing the tab or reloading the
page does not interrupt it.

The install chain is now shared: _ram_install_plan(body) builds the
card-independent part (the parsed .cap, the INSTALL/LOAD sequence, the
block size, the transport knobs) and _ram_install_run(server, body, plan,
scc) runs one chain (the per-card preset/keyset/counter resolution
included); the /api/ram-install handler is the plan + run + response, and
the batch worker calls the same pair.  The plan is built once at Start,
so a bad .cap or parameter combination is a 400 there (a malformed
archive included - a zip parse error is the caller's, never a dropped
connection), not a per-card failure.  _op_keyset_resolve() is the module
form of the handler's _op_keyset() (the dead _keyset_guard method is
gone).

The worker (_ram_batch_worker, state _RAM_BATCH) polls the card state
every 0.5 s and acts on a new card session while connected; a card
already equipped at Start installs immediately.  Per card the preset is
matched by ICCID and its ISD role TAR supplies the keyset (v3.26.0) and
the counter seed (v3.20.0); a card with no readable ICCID, no matching
preset or no usable ISD keyset is recorded as skipped with its reason.
A card is attempted once per run (any outcome): re-equipping it records
"already processed in this batch (<outcome>)" and never installs again -
restarting the batch is the retry.  Failures carry the failing step and
the remote SW, are never retried and never stop the run; an unexpected
card-handling error is recorded and the loop keeps waiting.

Endpoints: POST /api/ram-batch/start (no card needed; 400/409), GET
/api/ram-batch/status (card-free - the panel polls it during an install),
POST /api/ram-batch/stop (the first _CARD_FREE_POST: it only sets a flag,
so the button works while the worker holds the card lock; a running
install always finishes first) and POST /api/ram-batch/clear (409 while
active).  /api/status carries the ram_batch summary (the detailed state
minus the attempts list), so the panel's headline - the live phase, the
current card, the counts - rides the 2 s poll; the attempts report (the
last 500 kept, the total counted) is fetched only when it changed.  The
batch and a test run exclude each other (409 both ways); the state is in
memory - a server restart ends a run.

PWA: the Batch install (.cap file, per card) RAM operation shares the
install form (ramInstallCapFields(); the single install uses it too and
now sends ram_format, and the block-size validation moved before the
modal), hides the preset selector + Execute and the preset-mismatch
banner, and renders Start/Stop/Clear, the phase headline (equip the next
card / installing card + live step / stopping after the current card),
the counts and the attempts table (newest first).  The panel only works
inside the visible SCP80 RAM pill.

Docs: help EN/RU, api.md (a section + the four table rows), the repo
AGENTS.  Tests: the state/report helpers, the per-card decisions, the
worker loop (a real thread + a fake install), the lifecycle and guards,
the HTTP endpoints (the malformed .cap, the stop under a held card lock,
the test-run refusal) and the PWA panel helpers/wiring.  914 frontend +
919 Python; sw cache simple-v438.
This commit is contained in:
2026-10-14 15:00:03 +03:00
parent 21d7f58f61
commit ab373aff30
13 changed files with 1527 additions and 236 deletions
+1
View File
@@ -322,6 +322,7 @@
<tbody>
<tr class="border-b border-gray-200 dark:border-slate-700"><td class="py-1 px-2">Explore Card (all GP data)</td><td class="py-1 px-2">Queries GET STATUS for ISD, Applications, ELFs, and ELF Modules, plus GET DATA FF21 (Extended Card Resources: the number of installed applications and the free non-volatile/volatile memory). The compact listing carries the AID, life cycle and privileges (the legacy format has no version); if a card rejects it (<code class="font-mono text-sm">P2.b2=0</code>, deprecated by GP) the queries are retried in the expanded <code class="font-mono text-sm">E3</code> form, and the <strong>Fetch expanded registry details (load-file / SD associations)</strong> checkbox also requests it as an extra pass and merges the expanded-only fields: the Executable Load File version, Load File AID, Security Domain AID, module lists, the applet STK menu entries (SCP Registry Data, TS 102 226 8.2.1.6) and any other registry data objects the card returns (shown raw, never dropped) &mdash; it costs a few extra secured packets per Explore and is <strong>on by default</strong> (remembered per browser; untick to skip). Cards differ in what they accept, so the expanded query is tried in three forms &mdash; the reference tag list (with the TS 102 226 <code class="font-mono text-sm">EA</code>), the same list without <code class="font-mono text-sm">EA</code>, and no tag list at all (the GlobalPlatformPro form) &mdash; and the explorer reports which form worked and where the run fell back to the compact listing (package versions are then unavailable). When the ELF rows still carry no version, the tag-list forms are retried for both P1=10 and P1=20, skipping the combinations already refused, and the walk stops as soon as a version appears. Results appear in an explorer view (the success line stays minimal; the counts are in the section headings) with per-item <strong>Delete</strong> buttons and the detected RAM command format (hover for what compact vs expanded means).</td></tr>
<tr><td class="py-1 px-2">Install Package (.cap file)</td><td class="py-1 px-2">Sends a <code class="font-mono text-sm">.cap</code> file to the card via the server: INSTALL[for load] &rarr; LOAD &times;N &rarr; INSTALL[for install (+make selectable)]. The load file is split into LOAD APDUs of the <strong>LOAD block size</strong> (1&ndash;240 bytes of payload, 240 by default); each secured packet is delivered as a single SMS or, when larger, as a concatenated SMS-PP download of up to 5 segments, so a large <code class="font-mono text-sm">.cap</code> simply takes several SMS. As soon as the file is selected it is validated and its <strong>memory requirements are estimated</strong> in the form: the <strong>NVRAM requirement</strong> (the code image + persistent data, the GlobalPlatform Card Spec C6+C8 style total) and the RAM estimate. The same box reports the <strong>libraries the CAP is linked against</strong> (its Import component: library name, AID and the minimum version <code class="font-mono text-sm">≥ x.y</code>, with standard names and families resolved and a Java Card SDK hint where known), the package/applet identity and the component breakdown (sizes and shares in load-file order) &mdash; grouped under the <strong>Package</strong>, <strong>Requires</strong>, <strong>Memory</strong> and <strong>Components</strong> headings. The <strong>Platform</strong> line names the minimum Java Card level the card must provide (derived from the <code class="font-mono text-sm">javacard.framework</code> import; the header <code class="font-mono text-sm">int</code> flag additionally requires int support, Java Card 2.2+) and the <strong>ETSI release</strong> the UICC API imports come from (<code class="font-mono text-sm">uicc.toolkit</code> version &rarr; REL-N, per the workspace <code class="font-mono text-sm">sim-tools</code> corpus; shown with its basis in brackets): the JCRE checks every import against its own packages when it verifies the load file (JC VM spec 4.5.2), so a card whose platform is older rejects the LOAD as soon as the <strong>Import</strong> component is complete (the step line carries the card's remote SW and names the component the block completed, e.g. <em>completes Import</em>): the failure hint and the popup then point at the <strong>CAP compatibility test</strong> (one click &mdash; the same <code class="font-mono text-sm">.cap</code> is still selected) to find the missing package and version; a smaller LOAD block only helps when the block exceeds the card's buffer. A corrupt or wrong-format file fails the analysis, and <strong>Execute</strong> stays disabled until a successful analysis. The estimate is informational only and is never used for installation: NVRAM is the code image plus the persistent data (the GlobalPlatform C6+C8 style total, Table 11-48 &mdash; card memory management, rounding and the registry entry are not included, and the static field image is counted in both parts); the suggested <strong>C6</strong> follows the load file, and the suggested quotas are decimal byte counts (the hex in parentheses is how the C7/C8 install parameters encode them: C6 = non-volatile code, C7 = volatile RAM, C8 = non-volatile data).</td></tr>
<tr class="border-b border-gray-200 dark:border-slate-700"><td class="py-1 px-2">Batch install (.cap file, per card)</td><td class="py-1 px-2">The bulk workflow: one <code class="font-mono text-sm">.cap</code> and one set of install parameters, applied to card after card. <strong>Start batch</strong> needs no card &mdash; the server owns the loop: it waits for a card, installs the moment the equip finishes, then waits for the next one. Every card's <strong>preset is matched by ICCID</strong> and supplies the keyset (the ISD role TAR's default, v3.26.0), the keys and the counter; a card with no matching preset or no usable ISD keyset is reported and skipped. Each attempt lands in the list with its time, ICCID, preset, verdict and duration; a failure shows the failing step and the remote status word &mdash; the batch never retries and never stops on a failure (failures are yours to manage afterwards). A card is attempted <strong>once per run</strong>: re-equipping it reports &ldquo;already processed&rdquo; &mdash; restart the batch to retry it. <strong>Stop</strong> finishes the current card first (an interrupted LOAD chain can leave a half-loaded package) and then ends the run; <strong>Clear</strong> drops the finished report. The loop runs in the server, so closing the tab, switching views or reloading the page does not interrupt it; a server restart ends it. A batch and a test run exclude each other.</td></tr>
<tr class="border-b border-gray-200 dark:border-slate-700"><td class="py-1 px-2">CAP compatibility test</td><td class="py-1 px-2">Answers one question without installing anything: <strong>can this card satisfy the CAP's import list?</strong> The server runs the format check, INSTALL[for load] and the LOAD blocks only up to the block that completes the <strong>Import</strong> component &mdash; the point where the JCRE verifies the import list &mdash; then stops: no last-block flag and no INSTALL[for install], so nothing is committed and nothing has to be deleted afterwards. The <strong>Import versions to test</strong> list is pre-filled from the CAP analysis: <code class="font-mono text-sm">AID=version</code> rewrites an import's requested version, <code class="font-mono text-sm">+AID=version</code> appends a synthetic import (a card-capability query: the LOAD passes only if the card exports that package at &ge; that version), and <strong>Set all 0.0</strong> drops every requirement. Two-step use: press Set all 0.0 &mdash; if the test passes, the import list was the blocker; then restore one line at a time to find the package and version the card lacks. The verdict states it covers the <strong>LOAD/import gate only</strong>: the link gate (whether the applet's actual API calls exist) is what a real install checks at INSTALL[for install]. The run and its result appear in a popup (like the install): the step lines, the verdict and a per-AID detail block &mdash; on a rejection it names the CAP section the failing block covers (load-file byte range and components, and whether that block completes <strong>Import</strong>), the remote status word with its meaning, the import list exactly as tested (CAP versions, probe rewrites, appended synthetic entries, unmatched lines) and what the probe state implies; on success it lists the import list that passed.</td></tr>
<tr class="border-b border-gray-200 dark:border-slate-700"><td class="py-1 px-2">INSTALL [for install] (loaded package)</td><td class="py-1 px-2">One INSTALL [for install] APDU for a package that is already loaded on the card (its Load File AID and Module AID, privileges, install/STK parameters and an optional <em>make selectable</em>) &mdash; iterate the install parameters without deleting and re-loading the <code class="font-mono text-sm">.cap</code> every time.</td></tr>
<tr><td class="py-1 px-2">INSTALL [for make selectable]</td><td class="py-1 px-2">One INSTALL [for make selectable] APDU (the instance AID) for an INSTALLED, non-selectable applet instance.</td></tr>